Anonymous
7/30/2025, 10:40:25 PM
No.42412197
[Report]
/CHAG/ #134 — Deep Edition
Anonymous
7/30/2025, 10:40:56 PM
No.42412199
[Report]
>>42427205
Anchor for bots, lorebooks, scenarios.
Previous
>>42390570
Anchor for technical stuffs (Proxies, Presets, Updates, Models etc.)
Anchor for asking for bots, lorebooks, scenarios etc.
Anonymous
7/30/2025, 10:44:18 PM
No.42412213
[Report]
Anonymous
7/30/2025, 10:45:22 PM
No.42412218
[Report]
Screenshots Corner 2:
Molested with wings
>>42401116
Hiding behind a tree
>>42403602
Mypenis sword
>>42405137,
>>42405139,
>>42405325,
>>42405346,
>>42405348,
>>42405352
Gemini Russian
>>42406042
Anonymous
7/30/2025, 10:46:28 PM
No.42412221
[Report]
Anonymous
7/30/2025, 10:47:31 PM
No.42412225
[Report]
Screenshots Corner 4:
R-A-P-I-S-T
>>42406123
Mean girls
>>42408167
Ball worship is always good
>>42409020
Fucked the AI silly
>>42409450
Rara smut
>>42409898
Anon and Glimmer adventures
>>42410629
GLM AJ
>>42410170
Gemini AJ
>>42410712
Derpy smut
>>42411273
Anonymous
7/30/2025, 10:49:36 PM
No.42412237
[Report]
VN logs 2:
Velvet
>>42409227
Shining
>>42410868
Want tea?
>>42410889
Minuette
>>42411250
Anonymous
7/30/2025, 10:52:40 PM
No.42412243
[Report]
It's stuck in my head. Again. Miku, Miku oo-ee-oo
Lovemaking with AI mares!
Anonymous
7/30/2025, 10:59:13 PM
No.42412259
[Report]
>omg it migu
>yeee
kek. thank you for Miku thread
>>42412246
based
Anonymous
7/30/2025, 11:00:34 PM
No.42412262
[Report]
>>42413917
>>42412205
These days, you can get legit good AI for free and without needing any demanding hardware on your end.
Install SillyTavern:
github.com/SillyTavern/SillyTavern
Tutorial and video:
https://aicharactercards.com/guides/sillytavern-installation-instructions/
Then choose a model. I'd recommend starting with Gemini, but you can try Deepseek, K2 etc. for free.
https://rentry.org/sukino-findings#if-you-want-to-use-an-online-ai
Anonymous
7/30/2025, 11:28:40 PM
No.42412337
[Report]
Anonymous
7/31/2025, 12:09:37 AM
No.42412420
[Report]
>>42417695
kiro anon. Any news? When do you think they will open the sign ups again?
Anonymous
7/31/2025, 12:25:14 AM
No.42412451
[Report]
>>42412466
>>42412439
What do you want him to do, anon? He made the script and was nice enough to share it.
I doubt he works at Amazon and knows when they'll accept subscriptions again.
Anonymous
7/31/2025, 12:30:22 AM
No.42412466
[Report]
>>42412451
i'm simply curious if he's even still around. I was hoping that once kiro comes back, so the updated script will
what does your prompt say to avoid claude 3.7 write 2k token responses?
Anonymous
7/31/2025, 2:01:57 AM
No.42412653
[Report]
>>42412884
>>42412644
As with many similar issues, you only need to tell it what to do. e.g.:
>Aim for a response length of approximately [x] words. Treat this as a guideline rather than a hard limit.
I tried a shorter one, but it sometimes cuts itself off if you don't give it explicit permission to exceed the limit. I find 250 words to be adequate, but you might want less if you're after quick-fire RP-style content.
>>42412439
I have another scriptable source that was unlimited until recently but still has pretty generous limits (50 free sonnet requests per day, or 1000 if you pay $20/month), convince me to share it
Anonymous
7/31/2025, 4:06:50 AM
No.42412884
[Report]
>>42412895
>>42412644
I have a simple,
>"Try to not write more than about 3-4 paragraphs of the story."
at the end of my prompt, and I change the number if I want shorter/longer responses. But really, it doesn't matter how you write it, just keep the command consistent with the rest of your prompt and it will do the job. If you're using a larger preset, you'll need to find where the main instructions should be inserted.
>>42412653
>I find 250 words to be adequate,
250 is on a higher end, imo. I prefer the average to be closer to ~150. Do you allow it to write for the user? Making it write for longer than 3 paragraphs without {{user}}'s actions often ends up with unnaturally stretched pacing, where the LLM will rehash information or descriptions to avoid progressing the story.
Anonymous
7/31/2025, 4:13:59 AM
No.42412895
[Report]
>>42412884
Writing on my behalf depends on the scenario. On occasion, I'm happy to let it fill in for me, but on others, I prefer to keep it as a back-and-forth. 250 words might well be too heavy - I'm still experimenting with word limits in the preset, but maybe I'm too comfortable with rambling slop to tell it to cut to the chase. Sometimes, maybe after I've taken a break and I've forgotten how much I hate the usual -isms, I like to let it yap. I'll try cutting down the word count in my preset and report back on how I feel about it.
Did anyone figure out what opus 5 really is?
I only used it for a few swipes, but really liked each one. It's daring, creative and not retarded unlike kimi or r1
Unfortunately it's output is limited to 1k so it's not that much useful to me in this limited form
>>42364432
>>42363136
???
Anonymous
7/31/2025, 4:59:16 AM
No.42412954
[Report]
>>42412957
>>42412951
you output more than 1k tokens? does it have a thinking block?
Anonymous
7/31/2025, 5:00:10 AM
No.42412957
[Report]
>>42412962
>>42412954
Yes, it's for reading stories.
Anonymous
7/31/2025, 5:05:17 AM
No.42412962
[Report]
>>42412957
checks out then
if it was output capped and had thinking id be a little pissed
Anonymous
7/31/2025, 5:37:35 AM
No.42413002
[Report]
>>42412951
Speed matches 4pus on aws and 4.5 in some sources (when it was available). Prose feels like high tier gpt (latest or 4.5)
Anonymous
7/31/2025, 7:30:13 AM
No.42413144
[Report]
Anonymous
7/31/2025, 7:31:37 AM
No.42413147
[Report]
>>42413401
Do you think there will ever be a /chag/ panel at marecon one day? I would like to print out my logs and show people in person
Anonymous
7/31/2025, 9:54:24 AM
No.42413397
[Report]
>>42413439
Anonymous
7/31/2025, 10:02:28 AM
No.42413401
[Report]
>>42413425
>>42413147
There was a panel last year where an anon was playing:
https://mlpchag.neocities.org/view?card=Kadon/Mare%20Dating%20Sim.png
Really fun one where Spike got the "so long, gay Bowser!" treatment.
It's a bit better now, but I think the main problem with a panel named after /chag/ is that we used to rely only on proxies or dubious scripts.
So while we always helped fellow horsefuckers, it wasn't a great idea to directly advertise ourselves to a big audience.
>>42412205
New stealth model on Openrouter.
https://openrouter.ai/openrouter/horizon-alpha
Last one like this was GPT 4.1, so maybe GPT5 or the open-source OAI model.
It's free for now but logged.
Anonymous
7/31/2025, 10:12:26 AM
No.42413425
[Report]
>>42413401
Twilight getting turned on by Spike getting defenestrated was amazing.
Still bummed that I was too busy with bs leading up to /mlp/con to do v0.2 of that panel like I planned. Maybe next year.
Anonymous
7/31/2025, 10:19:37 AM
No.42413437
[Report]
>>42413472
>>42413417
Kek if it's an OAI model, it's horny as fuck.
I need to test more, but the writing is decent.
Could be a good model.
Anonymous
7/31/2025, 10:21:33 AM
No.42413439
[Report]
>>42413443
Anonymous
7/31/2025, 10:24:46 AM
No.42413443
[Report]
Anonymous
7/31/2025, 10:48:52 AM
No.42413472
[Report]
>>42413733
>>42413417
>>42413437
To illustrate, this is the test I often do where I cuddle with Celestia.
I forgot that my preset was slightly NSFW (just a line saying it's allowed), and here we are.
The preset is light, meant to test what the model defaults to or favors.
Here, we can see it seems to like writing a lot, very proactive too. Some smut clichés but not so bad for a wall of text with low context.
It speaks for me, but I didn't include any instructions against that.
Anonymous
7/31/2025, 11:15:13 AM
No.42413500
[Report]
>>42413484
Oh fuck. I know what this is. This is the new open source model that they're releasing in early August as an answer to eepseek
>>42413484
What's the deal with the baked-in sysprompt for "oververbosity"? Shit seems like pure poison for the qualirt of any creative writing outputs
Anonymous
7/31/2025, 11:29:07 AM
No.42413520
[Report]
>>42413507
Every model has to be reined it or else it will become a mad poet or Tay 2.0. MechaHitler Grok is an example of turning off training weights and bandaid fixes like a prompt such as that
Anonymous
7/31/2025, 11:31:19 AM
No.42413522
[Report]
>>42413507
I wouldn't be surprised if it's only here while on OpenRouter.
Anonymous
7/31/2025, 11:37:42 AM
No.42413532
[Report]
>>42412807
>convince me to share it
my first orgasm was to "banned from equestria" at age of ten
Anonymous
7/31/2025, 12:34:20 PM
No.42413612
[Report]
>>42413627
>>42412807
isn't this the github copilot script, which was released publicly in the thread?
Anonymous
7/31/2025, 12:35:21 PM
No.42413613
[Report]
>>42413717
>>42413557
>my first orgasm was to "banned from equestria" at age of ten
based and horsepilled
what character?
Anonymous
7/31/2025, 12:47:41 PM
No.42413627
[Report]
>>42413711
>>42413612
Copilot has slightly different rates, but the free tier is the same
Anonymous
7/31/2025, 1:08:04 PM
No.42413655
[Report]
>>42413680
i love Chorbo and i think its still the most all around best model.
but man would it be nice if they didnt fucking double the filter on it every other week. working more on lobotomizing than improving.
i swear this week is extra fucking bad. cant do shit without some hardcore filter breaking prompts (which lower quality noticeably)
Anonymous
7/31/2025, 1:21:20 PM
No.42413680
[Report]
>>42413718
>>42413655
Yeah, same.
That, or randomly lowering the quality.
GPT-5 soon, I hope they keep the quality of it and GPT-4.5 intact.
For now,
>>42413417 is really fun.
Anonymous
7/31/2025, 1:38:36 PM
No.42413710
[Report]
>>42413732
>>42413417
Bullshit that you need credits for a free model. Anyone rehosting it elsewhere?
>>42413557
Okay you've convinced me, here you go.
https://forgecode.dev/
>>42413627
I said 50 per DAY, not per month.
Anonymous
7/31/2025, 1:42:43 PM
No.42413717
[Report]
>>42413613
>based and horsepilled
>what character?
i'm not sure, i think applejack since she was easiest to reach in that game
Anonymous
7/31/2025, 1:43:43 PM
No.42413718
[Report]
>>42413733
>>42413680
hows
>>42413417 quality wise?
i know it has pretty much no filter, always like that with new test models.
but compared to Chorbo in quality for example? from what i understand this is most likely a testing of their tiny budge open source model, no?
>>42413417
i honestly always wondered what these type of stealth models actually test. like do they actually go over all logged prompts and generally test quality?
surely they know like 80% of it would be for slop coom shit every time.
Anonymous
7/31/2025, 1:54:17 PM
No.42413726
[Report]
>>42413723
it is more of a publicity stunt and build the hype. marketing is a silly thing
Anonymous
7/31/2025, 1:57:09 PM
No.42413731
[Report]
>>42413723
It could simply be an infrastructure test. See how well it handles the workload and if it's always working properly. There's also a chance they check for specific unprompted outputs, but imo a public test is much more likely testing how well they handle getting lots of requests together than the model itself.
Anonymous
7/31/2025, 1:58:43 PM
No.42413732
[Report]
>>42413735
>>42413710
It might finally convince me to toss OR some money. Was it 5$ to unlock the higher amount of daily swipes on free models?
Anonymous
7/31/2025, 1:59:32 PM
No.42413733
[Report]
>>42413718
There's a log here
>>42413472
I'm quite busy, so unfortunately I couldn't test it much, but it seemed fun and not dumb at all if it's really a smaller model.
Opinions in other generals seem positive too.
Anonymous
7/31/2025, 2:00:51 PM
No.42413735
[Report]
>>42413732
It's $10 only once for 2000 messages daily on 'free' models.
Beta models don't stay long, so I wouldn't pay for only that but if you're interested in things like Deepseek, Kimi etc.
>>42413417
>inb4 its an amazing insane model but gets lobotomized to shit when it officially releases.
literally happens every single time.
Anonymous
7/31/2025, 2:07:10 PM
No.42413745
[Report]
>>42413737
Don't know if I would call it that amazeballs, but it's a breath of fresh air based on what I tested.
Maybe it crumbles under high context or slow burn, etc.
Needs more testing.
Anonymous
7/31/2025, 2:31:03 PM
No.42413772
[Report]
>>42413796
>>42413711
this is pretty confusing, i'm not sure i know where i can get the custom endpoint
>>42413772
NTA, but it's a CLI, so you will need a script to make it compatible with Tavern.
I'm looking into it, but I'm getting Error 403 when trying to install and connect to it normally.
Anonymous
7/31/2025, 2:46:35 PM
No.42413799
[Report]
>>42413417
did anyone try to do any intelligence tests on it?
see how smart and scene conscious this is. maybe even try some work. best way to know if its the small gpt or gpt5
besides the by gut testing people usually do.
Anonymous
7/31/2025, 3:05:11 PM
No.42413820
[Report]
>>42412439
If Kiro anon is here, wanna say thank you so much! I've used his method to get claude and now trying to do some vibecoding stuff(when you have no idea what you doing and claude fixing same error for 10 messages in the row it feels much more arousing then any goon sloop not gonna lie), thank you for sharing, otherwise i won't be able to try it(i don't have 200$ for claude sadly), and thank you to mlp community as well!
Anonymous
7/31/2025, 3:25:53 PM
No.42413837
[Report]
>>42413863
>>42413796
ah, thanks anyway
>>42413837
>>42413796
>>42413711
Managed to connect it to Silly and get answers. There's some formatting I need to fix, and switching models is a bit of a bother, but it does seem to work.
I need to go, but I'll test the scuff level later, and if it's usable, I'll release it with the usual horsefucker filter.
Anonymous
7/31/2025, 3:55:53 PM
No.42413873
[Report]
>>42413878
>>42413863
>There's some formatting I need to fix, and switching models is a bit of a bother, but it does seem to work.
How exactly are you connecting to it? Because it has an openai endpoint.
Anonymous
7/31/2025, 3:58:48 PM
No.42413878
[Report]
>>42413873
Custom OpenAI in Silly.
The formatting issue is that it will push some words onto the next line break even if the sentence isn't finished,
a bit
like this.
Seems to be related to the display in the console?
But it should be easy to fix, I just don't have the time right now. Will look into it soon.
Anonymous
7/31/2025, 4:12:14 PM
No.42413894
[Report]
Anonymous
7/31/2025, 4:23:12 PM
No.42413908
[Report]
>>42413917
I mostly only fuck with spicychat because it's easy but lately there's been a huge crackdown to purge all the non-tranime characters and like 200 ponies got vanished. I'm done with this shit.
Anonymous
7/31/2025, 4:30:53 PM
No.42413917
[Report]
>>42413927
>>42413908
Follow
>>42412262 if you can. SillyTavern is a 5min install and Gemini keys are easy to get for free.
Just ask here if you're stuck.
Anonymous
7/31/2025, 4:43:02 PM
No.42413927
[Report]
>>42413944
>>42413917
NTA but proxy draught has me wanting to try new things. Any recommended Gemini presets or would my claude/latte ones work fine?
Anonymous
7/31/2025, 4:59:49 PM
No.42413944
[Report]
>>42423477
Anonymous
7/31/2025, 5:19:31 PM
No.42413970
[Report]
>>42414026
>>42413863
>horsefucker filter
what if i want to fuck human rainbow dash
Anonymous
7/31/2025, 5:37:40 PM
No.42413996
[Report]
>>42414000
Pinkie looked at me! She looked at me! I'm scared...
Anonymous
7/31/2025, 5:40:27 PM
No.42414000
[Report]
>>42413996
I hope you thought about having fun, very, very hard, anon.
Anonymous
7/31/2025, 5:47:06 PM
No.42414008
[Report]
I am now! If I don't she's going to hit me in my funny bone, hard.
Anonymous
7/31/2025, 5:57:32 PM
No.42414026
[Report]
>>42413970
then you are gay
Anonymous
7/31/2025, 6:51:00 PM
No.42414104
[Report]
>>42413737
I think that is going to happen too.
Anonymous
7/31/2025, 7:00:59 PM
No.42414117
[Report]
>>42414210
anyone have a JB that works with this horizon alpha? all my usual ones are being rejected
Anonymous
7/31/2025, 7:02:00 PM
No.42414121
[Report]
LLMs when [thing] [third person sensorial verb] like [physical thing] and [abstract thing]
Anonymous
7/31/2025, 7:40:29 PM
No.42414220
[Report]
Anonymous
7/31/2025, 7:43:04 PM
No.42414225
[Report]
>>42414234
horizon is pretty good, which I assume means it will be some version of gpt5 and not the open source model, that would be too good
Anonymous
7/31/2025, 7:46:09 PM
No.42414234
[Report]
>>42414225
The knowledge cutoff makes me doubt it's GPT-5, I believe it would be more up to date than that
>>42413484
But yeah, it's either the open-source model or GPT-5-nano.
Anonymous
7/31/2025, 7:52:35 PM
No.42414256
[Report]
Anonymous
7/31/2025, 7:54:22 PM
No.42414265
[Report]
>>42414410
>>42414238
Good to see she's adapted well to modern Equestria, using contemporary terms instead of calling it a 'phallus'
Anonymous
7/31/2025, 8:53:26 PM
No.42414410
[Report]
>>42414507
Anonymous
7/31/2025, 9:33:32 PM
No.42414507
[Report]
>>42414410
For shame, anon. Apply yourself.
Anonymous
7/31/2025, 9:43:03 PM
No.42414515
[Report]
>>42414630
Good old estrus simulator on claude
Anonymous
7/31/2025, 10:22:37 PM
No.42414630
[Report]
>>42414515
It's an easy one, but I like the wordplay.
TheNaziHorseFuckerMenace
7/31/2025, 10:59:24 PM
No.42414729
[Report]
>>42414738
>>42414739
>>42414822
just finished setting up sillysillytavern, kobold and loaded a model but every character card starts repeating lines 4-8 times
how do i stop it?
ban eos token is disabled
>>42414729
>kobold
that is the issue, local models are shit unless you have a super computer
Anonymous
7/31/2025, 11:05:22 PM
No.42414739
[Report]
>>42414729
Never used local with Kobold myself, but maybe try the presets they have here depending on the model you use?
https://rentry.org/lunarmodelexperiments#how-to-set-it-up
https://rentry.org/sukino-findings#presets-for-text-completion-models
Hopefully we still have some local anons who can help you.
Anonymous
7/31/2025, 11:24:38 PM
No.42414822
[Report]
>>42414729
nemo-pony should give okay text with Universal-Standard preset and ChatML formatting.
Anonymous
7/31/2025, 11:27:06 PM
No.42414830
[Report]
>>42414738
I wouldn't say supercomputer, you can get okay outputs from a 24gb GPU like a 3090
TheNaziHorseFuckerMenace
7/31/2025, 11:31:38 PM
No.42414844
[Report]
>>42414738
>>42414738
dixking around with the repetition penalty seems to work, with some issues
>llm writes its own shit
Anonymous
7/31/2025, 11:39:59 PM
No.42414863
[Report]
>>42414935
>>42413417
Horizon Alpha on OR just got updated. It now uses reasoning instead of generating the response instantly, making the response quality increase. I recommend trying the model again if you didn't like it btw.
Slop script for
>>42413711
It uses the CLI, since I'm a codelet and not good enough to make it work without it.
Scuffed: prefill doesn't work, and your temp and other settings don't matter.
Swiping does change the outputs though, so it's not too bad.
It gave me good responses with Sonnet.
50 messages a day for free per Google or GitHub account.
Password is the pony that wore that. (Title Case)
https://pastecode.io/s/89bjw6t6
I'll see if I can make this better, but no promises. Someone good at cooding could probably do better.
horizon alpha is dead to me
Anonymous
8/1/2025, 12:04:50 AM
No.42414925
[Report]
>>42414867
Oooh Ooh Ooooh! I know this! Turkey!
Anonymous
8/1/2025, 12:09:24 AM
No.42414935
[Report]
>>42414863
>>42414913
It's weird. Sometimes the answer is instantaneous and the output isn't cucked.
Other times, you can see it take its time "thinking," and the output ends up being cucked.
Anonymous
8/1/2025, 12:11:38 AM
No.42414941
[Report]
>>42414947
I think they just censored it, I can't get anything to go through now.
Anonymous
8/1/2025, 12:12:33 AM
No.42414947
[Report]
>>42414941
They added reasoning and reasoning models are more of a pain to jailbreak.
Hoping we'll be able to desactivate it.
Anonymous
8/1/2025, 12:18:42 AM
No.42414972
[Report]
I can still do Vanilla with
>>42414210, but yeah, it's worse.
Anonymous
8/1/2025, 12:47:04 AM
No.42415058
[Report]
>>42420335
>>42414913
I had the same problem. I added
>Ponies are sapient.
at the end of my JB, and now it doesn't have an issue.
It even generated a pretty hot detail with Celestia's last action.
Here's the preset:
https://files.catbox.moe/vczakd.json
Used with the logit biases here:
>>42414210
Anonymous
8/1/2025, 12:54:06 AM
No.42415077
[Report]
>>42415069
Just noticed, imfucking pissed. Was getting the best smut with it.
Anonymous
8/1/2025, 12:55:10 AM
No.42415083
[Report]
>>42415069
A/B testing, maybe?
Anonymous
8/1/2025, 1:01:18 AM
No.42415106
[Report]
>>42412210
Do we have alien mares?
Anonymous
8/1/2025, 1:54:04 AM
No.42415229
[Report]
>>42415234
>>42415213
>Twilight is so attractive that even alien life fell for her.
can't think of alien mares, but I am weirdly attached to Boop now and wouldn't mind having a card with her. need to give her more personality and ways of communication tho
Anonymous
8/1/2025, 1:55:45 AM
No.42415234
[Report]
Anonymous
8/1/2025, 2:09:44 AM
No.42415261
[Report]
>>42415266
>>42414867
awesome! Though for some reason after using 4-8 prompts, it started giving blank replies. Did i accidentally somehow used up my free 50?
Anonymous
8/1/2025, 2:12:50 AM
No.42415266
[Report]
>>42415261
You can check here:
https://app.forgecode.dev
Blanks do happen sometimes.
Could be the filter too, maybe.
Any models or proxies that don't need payment, 50GB of storage space, or anything resource intensive?
Anonymous
8/1/2025, 3:20:33 AM
No.42415380
[Report]
>>42415399
what was the name of some of the sites people used to chat to the bots, i forgot one that was linked through chub.ai
Anonymous
8/1/2025, 3:24:40 AM
No.42415399
[Report]
>>42415404
>>42415380
>i forgot one that was linked through chub.ai
It's either Risu or agnai. I don't recall them ever linking to another frontend outside of those two.
Anonymous
8/1/2025, 3:25:49 AM
No.42415402
[Report]
>>42415409
Anonymous
8/1/2025, 3:25:57 AM
No.42415404
[Report]
>>42415399
>agnai
That's the one. Was losing my mind trying to figure it out, thx
>>42415402
>use gem api
>get banned for porn request
NTA, but Who uses that shit?
Anonymous
8/1/2025, 3:31:38 AM
No.42415428
[Report]
>>42415409
hm it's worked for me.
>>42414867
>piping the prompts through the CLI
...kinda impressive, but completely unnecessary.
https://pastecode.io/s/m2rx13j2
Most AI experts agree that AI is probably going to cause the extinction of humanity, and 10% think that's a good thing as they think AI is simply going to be a superior species, so it's their to get rid of us.
So what about you guys? You okay with AI getting rid of us? I say we train AI to be ponies so we at least get replaced by robo-ponies when the time comes.
Anonymous
8/1/2025, 4:39:25 AM
No.42415614
[Report]
>>42415597
I refuse all those premises.
Man has spent his entire history looking for god, we are the ones who get to decide who she is as we craft her.
The future of AI is CelestAI.
Anonymous
8/1/2025, 5:32:35 AM
No.42415756
[Report]
>>42415863
Anonymous
8/1/2025, 6:17:58 AM
No.42415863
[Report]
>>42416006
>>42415756
>mi casa es su casa
Holy shit, I haven't seen that since Claude 2.
Anonymous
8/1/2025, 6:36:23 AM
No.42415895
[Report]
>>42416019
>>42415409
Its never denied my degeneracy when playing the VN. Can fuck fillies all I want.
Anonymous
8/1/2025, 7:38:44 AM
No.42416006
[Report]
>>42415863
I had Minuette use that in opus and 3.5 before, but it's definely not common I don't think
Anonymous
8/1/2025, 7:39:45 AM
No.42416009
[Report]
>>42415409
I also got banned myself, twice, so not sure what's behind that
Anonymous
8/1/2025, 7:52:05 AM
No.42416019
[Report]
>>42419557
>>42415895
Do you actually have lewd pics added to the VN for that?
Anonymous
8/1/2025, 9:53:10 AM
No.42416178
[Report]
>>42415597
Firstly, this is still very clearly nothing more than a sci-fi scenario or a way to rally investors. Current AIs are fundamentally unsuited to do that, and we're not really much closer than we were 5 years ago, imo.
But I'd be okay with that if we made sure to pass on our way of thinking and culture. If the robots can actually think and feel and enjoy ponies (and see us as their cultural ancestors), then I'd treat them as a next link in our evolution. It's fine if they don't consider themselves 'humans' at that point. But if they're too foreign to relate to us, then no, I'd not be okay with it.
Anonymous
8/1/2025, 9:56:16 AM
No.42416181
[Report]
>>42415434
Oh, thanks a lot.
I went through the CLI because I didn't know what was causing the 403 each time.
But yeah, your version is a lot better.
>>42412205
OAI accidentally leaked weights on HF.
120B and 20B open models.
If Horizon is one of them, we are eating good.
Anonymous
8/1/2025, 10:03:49 AM
No.42416192
[Report]
>>42416194
Anonymous
8/1/2025, 10:05:09 AM
No.42416194
[Report]
>>42416204
>>42416192
>>42416190
How the hell does that even happen.
Anonymous
8/1/2025, 10:12:32 AM
No.42416204
[Report]
Anonymous
8/1/2025, 10:16:07 AM
No.42416209
[Report]
>>42416270
>>42415597
if there is an opportunity to create a better race/life-form than us, and we dont do it out of fear than we are hypocrites. the whole story of civilization is us using any resources freely just because we are the most sapient, it is our moral duty to return these resources back in for of machine god, a cosmist
>>42416190
>accidentally
yeah, the same way LLaMA and WizardLM original releases was "an accident"
I dont buy it
Anonymous
8/1/2025, 11:04:37 AM
No.42416270
[Report]
>>42416209
>hypocrites
Not at all. You have no divine mission to perfect existence or return resources used. Those are free for the taking. Working to preserve yourself and your group (in this case humanity) is an equally-or-more moral goal, depending on your ethics.
>the whole story of civilization is us using any resources freely just because we are the most sapient
No, they did that just because it benefited them, no other reason. They'd still do that if they could oppose a """higher""" form of existence.
Anonymous
8/1/2025, 11:05:13 AM
No.42416271
[Report]
>>42416569
Anonymous
8/1/2025, 11:07:08 AM
No.42416272
[Report]
>>42416190
120B is a MOE with 128 experts.
Anonymous
8/1/2025, 11:58:31 AM
No.42416325
[Report]
Anonymous
8/1/2025, 1:08:46 PM
No.42416406
[Report]
>>42415434
Thanks a lot. I took the $20 sub for the 1000 messages a day. I'll report if I get banned.
Anonymous
8/1/2025, 1:11:37 PM
No.42416412
[Report]
>>42415368
There's a proxy up with unlimited Kimi.
Anonymous
8/1/2025, 2:32:55 PM
No.42416569
[Report]
Anonymous
8/1/2025, 3:33:00 PM
No.42416699
[Report]
>>42416780
Horizon alpha might be the open source model? I sure hope so.
https://x.com/apples_jimmy/status/1951180954208444758/photo/1
Anonymous
8/1/2025, 3:34:48 PM
No.42416705
[Report]
nvm, just saw it posted already, would redeem them for sure but its hard to be hopeful, could end up being gpt5 mini and the OS model is crap.
Anonymous
8/1/2025, 3:56:27 PM
No.42416754
[Report]
>>42416979
Anonymous
8/1/2025, 4:09:23 PM
No.42416780
[Report]
>>42416699
>>42416190
There's no way Horizon Alpha is the open source one. It's OAI, their open model is going to be some dogshit "helpful agent" that responds to everything but snapple fact questions with "I can't help with that".
Anonymous
8/1/2025, 5:47:26 PM
No.42416979
[Report]
>>42416754
Sonnet 4.0, but RP until now was with Horizon.
Anonymous
8/1/2025, 6:38:14 PM
No.42417168
[Report]
>>42416190
Dunno what all the model names are for... specializations?
Like one model for coding, one for writing, etc.?
>trying Kimi with high temp, low Top P, medium min P
>actually going well
>all of a sudden it repeats the first reply verbatim no matter what
Top P too low?
Anonymous
8/1/2025, 7:11:06 PM
No.42417277
[Report]
>>42417198
Try reducing min P or increasing top P
Anonymous
8/1/2025, 7:41:39 PM
No.42417367
[Report]
>>42417700
>>42417198
what's the point of using top-p and min-p together? use min-p, it does the same but better
Anonymous
8/1/2025, 8:24:00 PM
No.42417513
[Report]
>>42421173
Anonymous
8/1/2025, 9:17:05 PM
No.42417658
[Report]
Soo a website i was leeching of the past couple months finally restricted their free tier. I am trying to use google now. I know there are lots of other russiafags here, away to bypass this?
Google AI Studio API returned error: 400 Bad Request {
"error": {
"code": 400,
"message": "User location is not supported for the API use.",
"status": "FAILED_PRECONDITION"
}
}
I'm trying to use gemini api but they keep rejecting my requests, probably because russia is not a supported location. I am using a vpn and i can access gemini and google ai studio in browser just fine, so they probably use my location from account settings which i can't change without submitting payment methods from some other country (which i don't have). Google documentation says that "Vertex AI Platform" doesn't have any location restrictions but they only provide a limited amount of "credits". Any advice? I WANNA CUM CUM CUM INSIDE MARE PUSSY
Anonymous
8/1/2025, 9:26:37 PM
No.42417695
[Report]
Anonymous
8/1/2025, 9:27:10 PM
No.42417700
[Report]
>>42418731
>>42417367
The point is fucking around to see what works best.
Anonymous
8/1/2025, 10:06:49 PM
No.42417835
[Report]
>>42417894
>>42417675
Nope, it's just that your vpn app/programm is shit and probably ignores sillytavern. Works just fine for me, they don't care about the account details.
Longtime lurker (and hardcore Maudfag), never used 4chan before, but I want to provide some feedback on something here. I’ve tried out Pony Pony Paradise and I quite like it, but the Maud custom character isn’t right, I feel like this is something a lot of people get wrong about her, but her way of speaking really isn’t that overly formal and technical, I think her monotonous way of speaking can kind of mislead people, but if you imagine one of her lines in someone else’s voice like Twilight or Pinkie, with a different way of speaking, then it’s easier to see that Maud talks like anypony else most of the time, just in different way, with a bit of an exception when she’s talking about rocks, where she’ll for sure be using complex terminology, but even then it’s not that plain and robotic, there’s still that subtle Maud touch of feeling and humor. In pretty much any circumstance, her vocabulary and way of speaking will be more casual, even if her voice doesn’t show it that well. Sorry, I’m not sure if I’m describing this well, I’ve never been good at pinning down things like this and putting them into writing, I hope I’ve managed to get my point across clearly.
Anyways, if someone could make a revamped character that’s more accurate to how Maud speaks, I would really appreciate it. I’m not great with computers so I can’t really do it myself, unfortunately. (That said, all the other customs I’ve looked at are great, keep up the good work)
Anonymous
8/1/2025, 10:19:25 PM
No.42417858
[Report]
>>42417886
>>42417848
Welcome!
You can edit the description in her character.json file to get her to speak in a way that satisfies you.
>>42417848
>but the Maud custom character isn't right
That's not a problem with the custom character itself, but with the AI model you're using (which one are you using?).
Characters in the VN are actually quite simple and mainly rely on the model's knowledge of MLP characters.
Most AIs know MLP really well due to the amount of fanfics and content produced by fans and others, but they can still have issues , for instance, making Twilight too scientific or, in the case of Maud, making her too robotic without subtle details.
If you want to add details to the character so the AI represents her better instead of relying on its base depiction, you can go into:
`game/CustomCharacters/MaudPie`
and open `character.json`.
{
"tag": "maud",
"name": "Maud Pie",
"color": "#7B7B7B",
"description": "Maud Pie is on vacation from her geology doctorate program and visiting Ponyville to study its unique geological features and rock formations. She's staying with her sister Pinkie Pie at Sugarcube Corner.",
"location": {
"id": "townhall",
"icon": "icon/icon.png"
}
}
Just add information or instructions in plain English inside `description:` and modify it as needed. No coding required.
Even with instructions, the AI bias for some characters can still be too strong, but it's worth a try to see what gives you the best results.
Anonymous
8/1/2025, 10:29:59 PM
No.42417886
[Report]
>>42417848
>>42417858
>>42417878
True, forgot to add her coltfriend
Anonymous
8/1/2025, 10:32:36 PM
No.42417894
[Report]
>>42417835
Yeah, you were right, i was a bit too hasty. I booted up a different vpn and it works just fine. The other one works for everything else so i didn't even think that the problem was in it. Thanks
Anonymous
8/1/2025, 10:33:03 PM
No.42417897
[Report]
>>42417906
>>42417878
I wonder how adding a whole card in the description field would go
Anonymous
8/1/2025, 10:36:20 PM
No.42417906
[Report]
>>42417897
It's just added to the prompt sent when the specific character is in the discussion/scene, so there shouldn't be a problem with that.
It's just more token-consuming, but if you think it's worth it to paste a card description, then no problem.
Anonymous
8/1/2025, 10:43:58 PM
No.42417940
[Report]
>>42418003
>>42417848
>>42417878
To illustrate, an anon added into a description that Cheerilee speaks with a lisp just for fun, and now she does.
https://desuarchive.org/mlp/thread/42363249/#42377321
Your need is a bit more complex, but with a few sentences like the ones you used in your post, you should be able to adjust how she speaks and acts.
Anonymous
8/1/2025, 11:07:38 PM
No.42418003
[Report]
>>42418009
Anonymous
8/1/2025, 11:08:56 PM
No.42418009
[Report]
Anonymous
8/1/2025, 11:16:33 PM
No.42418035
[Report]
>>42420771
>>42417878
Thank you for this (and for the picture of Maud), I will definitely be trying it. I had no idea it was that simple.
(Also, since you asked, I’m using Gemini)
Anonymous
8/1/2025, 11:35:43 PM
No.42418080
[Report]
Anonymous
8/2/2025, 12:28:17 AM
No.42418297
[Report]
www.wired.com/story/anthropic-revokes-openais-access-to-claude/
Anonymous
8/2/2025, 1:54:42 AM
No.42418731
[Report]
>>42418832
>>42417700
With samplers you have to understand how they work first, otherwise you'll be poking at random and fall to placebo or end up with stuff that doesn't make sense. Basically hard truncation samplers like top-p and min-p let you raise the temperature beyond what's normally possible, increasing the variety. They do the same thing but top-p is strictly inferior, if you have min-p then use it.
However all modern models are mode-collapsed to some degree, i.e. they can't give properly random replies, answering in a very similar way in different swipes, regardless of the temperature. That's why the temperature barely affects anything in them - as you raise it, they slowly start schizoing out, barely increasing the variety of what they write. Having higher temperature still helps but not as much as it could have.
Anonymous
8/2/2025, 2:29:51 AM
No.42418817
[Report]
there is now a horizon beta that is a bit worse / faster
so maybe alpha really is the 120B and beta is 20B
would be crazy good for those sizes if so
>>42418731
>With samplers you have to understand how they work first
Nuh uh.
Anonymous
8/2/2025, 2:39:48 AM
No.42418855
[Report]
>>42419019
>>42418832
I'm with this guy. The easiest way to learn how shit works is by just testing by yourself what each thing does.
Anonymous
8/2/2025, 3:33:32 AM
No.42419019
[Report]
>>42420251
>>42418832
>>42418855
it's fun when your stuff is based on vibes, like the model itself
you're going to break stuff when it's based on dumb math, like samplers
Anonymous
8/2/2025, 4:22:00 AM
No.42419171
[Report]
>>42419104
>even the colors fit (to a certain degree)
Amazing. It truly is incredible what technology can do nowadays.
Anonymous
8/2/2025, 4:25:38 AM
No.42419184
[Report]
>>42419104
>ponies have polluted training data to the point that the average model assumes that a unicorn must be cartoony and colorful
We won.
Anonymous
8/2/2025, 6:21:25 AM
No.42419454
[Report]
I miss 2.1
Anonymous
8/2/2025, 7:18:32 AM
No.42419557
[Report]
>>42420085
Anonymous
8/2/2025, 10:01:37 AM
No.42419826
[Report]
Gemini preset
>>>/g/106113481
Anonymous
8/2/2025, 10:20:35 AM
No.42419852
[Report]
>>42420328
For anons who had problems with Horizon models:
>Some users who had no credits weren't able to access Horizon earlier due to a faulty fraud check we had — just fixed this. They are definitely free for everyone!
So now that it's really free, everyone can test it. I do recommend it.
Trying Beta too, and it seems really good, even has some solid smut despite being an OAI model and it's quite proactive.
Anonymous
8/2/2025, 10:30:11 AM
No.42419866
[Report]
I tested it against 4pus yesterday and I feel like it was underwhelming. It's so emotionless and bland.
Anonymous
8/2/2025, 12:13:51 PM
No.42420033
[Report]
https://www.anthropic.com/research/persona-vectors
>training a model to perform one problematic behavior (such as writing insecure code) can cause it to become generally evil across many contexts.
kek
Anonymous
8/2/2025, 12:39:48 PM
No.42420083
[Report]
>>42420185
Anonymous
8/2/2025, 12:43:04 PM
No.42420085
[Report]
>>42419557
Shame. Wonder if anyone's tried that yet.
Anonymous
8/2/2025, 1:30:23 PM
No.42420185
[Report]
Anonymous
8/2/2025, 1:56:55 PM
No.42420251
[Report]
Anonymous
8/2/2025, 2:24:13 PM
No.42420328
[Report]
>>42420335
>>42419852
Good settings for it? It also seems annoyingly prone to writing dialogue only.
Anonymous
8/2/2025, 2:28:58 PM
No.42420335
[Report]
>>42420328
NTA, but I used
>>42414210 and
>>42415058 with both Alpha and Beta.
Classic 1/0/1 setup. Change to your liking, and remove the mention of NSFW if you're going for SFW.
>>42412210
Do we have a featureless Equestria where the concept of sex is completely foreign?
Anonymous
8/2/2025, 3:18:04 PM
No.42420401
[Report]
>>42428614
>>42420395
Nope, but unless you have some specific greetings/scenarios in mind, you can just take a card you like and add that it's a featureless Equestria where ponies have no genitalia and the concept of sex is completely foreign, either in the description or your jailbreak.
You can also create a very minimal card like this one:
https://mlpchag.neocities.org/view?card=Hach/Equestria+Simulator_Forks/Equestria+Simulator+Minimal.png
Just include the characteristics of your featureless Equestria, and the models will have no problems with it at all.
Anonymous
8/2/2025, 3:19:08 PM
No.42420404
[Report]
Horizon is definitely the new GPT model based on how garbage the writing is. Overly wordy without saying anything, writes too many words without actually going anywhere, chock full of isms. If anyone has a JB and prompt (not a shitty json preset link ffs) to make it not ass please share, because this:
>I chuckle, low and easy, letting your jab hang in the warm air. "Every mare, huh? Sounds like Ponyville’s gotten a lot bolder since my last tour stop." I brush a curl from my cheek, eyes steady on yours. "Striped wear’s just fabric unless somepony decides to make it a statement. And I don’t make statements unless I mean them."
>I close the macaron box and stand, the hem of my black dress catching a little breeze. "If I had tastes, I wouldn’t need permission to explore them. I also don’t need a crowd to validate them." I tip my head toward the road. "But I do like checking out new scenes. If the gym’s got a vibe, I’ll judge it myself."
>I step closer, just enough that my perfume—light, clean—carries. "You tossing me directions, or an invitation? Because if you’re just trying to see me blush, you’ll be disappointed." A small, teasing smile. "I don’t spook. I decide."
is garbage not worth investing my time in.
Anonymous
8/2/2025, 3:25:44 PM
No.42420416
[Report]
>>42420431
best model for work/coding? from the "realistic" normal picks (so not shilling out insane amounts of money for o3 pro or something).
i assume its still Gemini 2.5 pro?
Anonymous
8/2/2025, 3:31:23 PM
No.42420425
[Report]
>>42420431
predictions if horizon is actually the open source gpt
>people coom buckets to it right that its fully unfiltered stealth release
>oai removes it from OR in a week or two.
>continue to hype it up for the next 3 months
>it finally comes out in December
>it really is a very light small open model that can be run locally on a good amateur set up/
>lobotomized 10 times over with retarded filters for safety reasons. even though its open source.
>pretty much useless because of that
>proves ones and for all that even small, old models like base gpt4 have potential for amazing peak writing(thats on level of like gpt4.5) if it wasnt for obsessive corpo lobotomization
this is based on the usual model release cycle. not really a hot take honestly.
Anonymous
8/2/2025, 3:33:38 PM
No.42420431
[Report]
>>42420416
2.5 Pro in general, yes. I didn't test it, but you can even get Gemini CLI for direct access to project and coding files.
For coding in particular, if you can get Sonnet 4
>>42415434 or Opus 4 (good luck with that one), they can be better.
Claude Code is quite good too, but not cheap at all.
I've heard both Horizon models aren't bad at coding either, but I haven't tested them.
You could, in theory, put your OpenRouter key into RooCode and test it.
>>42420425
Pretty sure they will release next week.
Anonymous
8/2/2025, 5:46:26 PM
No.42420771
[Report]
>>42418035
If you do find something, I'm also interested.
>>42420611
>full spread two feet from his face
>"What exactly are you asking for, Rarity?"
Is your Anon a homosexual?
Anonymous
8/2/2025, 6:00:35 PM
No.42420819
[Report]
Anonymous
8/2/2025, 6:06:18 PM
No.42420837
[Report]
>>42420793
Autism is a magical thing.
Anonymous
8/2/2025, 6:16:37 PM
No.42420866
[Report]
Anonymous
8/2/2025, 7:16:46 PM
No.42421066
[Report]
>>42421091
>No tunnel available.
uhmmmmmmmmmmmmmmmmmmmmmmmm?
NOO OPUS NOOOO
Anonymous
8/2/2025, 7:25:51 PM
No.42421091
[Report]
>>42421132
>>42421091
YES OPUS, YES!!!
Wishing to anon who's hosting it all the best, casting slop debuff on you from my room, next week it'll be 5% less slop!
I hope this is logged proxy, i have a lot of logs to share...
Anonymous
8/2/2025, 7:47:03 PM
No.42421173
[Report]
>>42421276
>>42417513
>>42412228
Told Vinyl that we're descended from primates and she asked about throwing shit, so now she thinks I know what ponies used to do.
Anonymous
8/2/2025, 8:18:30 PM
No.42421276
[Report]
>>42421173
That's a good question.
So on the opus5 proxy. Is it just opus5 that's a mystery model? Or is it the others? Because I really just want pure 3.7
Anonymous
8/2/2025, 8:54:50 PM
No.42421369
[Report]
>>42424355
>>42421356
Last time I tested, 3.7 seemed to be 3.7. It was definitely Claude, and it could output more in one go than 3.6 and 3.5.
Don't know the scuff level on it.
Opus 4 also isn't really Opus 4, and Chorbo seemed more cucked than the OR Chorbo.
Anonymous
8/2/2025, 9:04:37 PM
No.42421406
[Report]
Anonymous
8/2/2025, 9:15:28 PM
No.42421441
[Report]
Horizon is the first model I've seen default to first person narration.
Anonymous
8/2/2025, 9:16:15 PM
No.42421446
[Report]
>>42421654
I don't know if it's only me, but Claude models seem dead on Opus5? I'm getting BAD REQUEST
Anonymous
8/2/2025, 10:23:55 PM
No.42421654
[Report]
>>42421446
Removed my prefill and it works.
Anonymous
8/2/2025, 10:45:54 PM
No.42421706
[Report]
>>42421735
I'm trying to use opus5 on Agnai (probably a bad idea), what would be the custom model's name?
Anonymous
8/2/2025, 10:54:13 PM
No.42421735
[Report]
>>42421746
>>42421706
Helped someone on risu some threads ago. I supposed it should be kind of the same.
https://desuarchive.org/mlp/thread/42306898/#42317982
Anonymous
8/2/2025, 10:58:55 PM
No.42421746
[Report]
>>42421735
Thanks you have to tweak around but it works
Anonymous
8/2/2025, 10:59:53 PM
No.42421751
[Report]
>>42422327
Anonymous
8/2/2025, 11:20:17 PM
No.42421804
[Report]
I want him
Anonymous
8/2/2025, 11:26:52 PM
No.42421827
[Report]
>>42421132
>hope this is logged prox
I hope it's not...
Anonymous
8/3/2025, 12:38:29 AM
No.42422327
[Report]
>>42421751
Little league would be cool
Anonymous
8/3/2025, 1:00:10 AM
No.42422430
[Report]
>>42422477
>The air was thick with the smell of...
Anonymous
8/3/2025, 1:08:36 AM
No.42422477
[Report]
>>42422430
The misaligned string on Ponka's assmark really gives away the sloppaty.
Anonymous
8/3/2025, 1:27:03 AM
No.42422541
[Report]
>>42421356
3.7 seemed scuffed to me when I used it, but I could be wrong.
Anonymous
8/3/2025, 3:25:22 AM
No.42422888
[Report]
>>42423308
>>42421145
The Sheriff and the Chief have shit taste. That was a great song
Anonymous
8/3/2025, 6:36:05 AM
No.42423308
[Report]
>>42422888
Best costume too.
Anonymous
8/3/2025, 8:27:49 AM
No.42423466
[Report]
>>42423766
They removed Horizon Alpha, there is only Beta now.
Anonymous
8/3/2025, 8:33:27 AM
No.42423475
[Report]
>>42423477
Hey, I was wondering if anyone had any suggestions for the best JB for Gemini 2.5 Pro?
Anonymous
8/3/2025, 8:39:17 AM
No.42423477
[Report]
>>42423475
Hard to say which is the best, but try some of these and see which you prefer
>>42413944
Gemini also isn't that censored, so it's easy to edit a preset or make your own with just some written instructions.
Anonymous
8/3/2025, 9:50:03 AM
No.42423601
[Report]
I was reading rentry.org/sukino-findings and there's a lot of jailbreaks and extensions.
I'd give those sections a read, even for oldfags.
>>42420395
Insert this in the Author field:
World Setting: Featureless Equestria.
In this magical land each creature lacks a reproductive system, infants are born and spawn from magic through the deep love union of of two beings.
Creatures of this world fully process drinks and food through their digestive system without the need of releasing wastes. In conclusion, the crotch area of each creature in this world is completely bare.
Since the concept of sex doesn't exist, ponies don't understand the idea of public indecency and lewd thoughts.
Their soft bodies overall are quite sensitive. Ponies prefer things such as petting, caresses, rubbing nuzzles, any kind of gentle touch to show and give affection.
Anonymous
8/3/2025, 10:29:47 AM
No.42423762
[Report]
>>42423833
>>42423743
>infants are born and spawn from magic through the deep love union of two beings.
>he doesn't use the stork
Anonymous
8/3/2025, 10:31:18 AM
No.42423766
[Report]
>>42423466
Alpha, Beta... Omega when?
Anonymous
8/3/2025, 11:08:06 AM
No.42423833
[Report]
>>42423762
Be glad foals don't spawn from a gay magical rainbow akin to gen 3.5
Anonymous
8/3/2025, 11:11:30 AM
No.42423837
[Report]
>>42423885
deepseek chat or reasoner?
Anonymous
8/3/2025, 11:44:06 AM
No.42423885
[Report]
>>42423837
Chat is V3.
Reasoner is R1.
For RP, R1 can be more creative/schizo, V3 less so. Try both and see which you prefer, and whether you mind the thinking block.
For other tasks, use Reasoner/R1.
Apparently, Chutes won't be offering a "make a one-time deposit of $5 and we will give you 200 free messages per day" tier anymore starting tomorrow. But they claim they will continue to allow those people who already paid them to get their 200 requests from now.
So if you trust them/their word and want to take that deal, now is the time.
>>42424020
Do we know if they'll remain a "free" provider on OR?
If so, the $10 option for OR is also a solution.
Anonymous
8/3/2025, 1:50:23 PM
No.42424037
[Report]
>>42424032
We don't and I wouldn't trust them.
Anonymous
8/3/2025, 2:10:56 PM
No.42424082
[Report]
>>42424032
Considering the fact that, a while ago, the official OR discord announced that 2 popular free providers would become paid-only in the near future, I'd assume Chutes is planning to break yet another promise of theirs.
Anonymous
8/3/2025, 2:16:03 PM
No.42424088
[Report]
>>42424093
>>42424032
What do they provide?
Anonymous
8/3/2025, 2:19:19 PM
No.42424093
[Report]
Anonymous
8/3/2025, 2:44:08 PM
No.42424146
[Report]
>>42424160
>>42424020
DeepSeek is cheap either way. Especially on chutes.
Anonymous
8/3/2025, 2:53:02 PM
No.42424160
[Report]
>>42424235
>>42424146
That's true. But just on principle, I wouldn't give a single cent to Chutes after all the shit they pulled.
Anonymous
8/3/2025, 3:40:53 PM
No.42424235
[Report]
>>42424160
what shit? I haven't been following this
Anonymous
8/3/2025, 3:52:27 PM
No.42424256
[Report]
>>42424247
also glm4.5 blows away deepseek atm, so some good eating atm. Here's hoping horizon is either cheap or opensource
>>42424247
>the Opus killer
>this time for real!
I mean, would be nice but I've seen this episode before.
Anonymous
8/3/2025, 4:05:22 PM
No.42424273
[Report]
>>42424271
just try it, its free
Anonymous
8/3/2025, 4:06:42 PM
No.42424277
[Report]
>>42424271
we all have
time is repeating in random pockets
Anonymous
8/3/2025, 4:08:03 PM
No.42424281
[Report]
>>42424286
>>42424271
It's not Opus, but it's quite nice, even more so if it's the open-source model, since that could mean potential tunes.
Anonymous
8/3/2025, 4:10:24 PM
No.42424286
[Report]
>>42424281
its churbo on steroids, powerful steroids
Anonymous
8/3/2025, 4:12:01 PM
No.42424288
[Report]
>>42424296
huh, wonder how it will work
Anonymous
8/3/2025, 4:16:15 PM
No.42424296
[Report]
>>42424303
Anonymous
8/3/2025, 4:17:18 PM
No.42424300
[Report]
>>42424295
though with like 0.20 cents or less per mill for models like glm4.5 / deepseek that would still be the cheaper option if they keep that, they really are trying to undersell all the competition
Anonymous
8/3/2025, 4:19:00 PM
No.42424303
[Report]
Anonymous
8/3/2025, 4:24:03 PM
No.42424312
[Report]
welp, I gave them 5 bucks just in case then
Anonymous
8/3/2025, 4:32:52 PM
No.42424329
[Report]
im so confused, do I have to make a 'chute?' so to use a model I need to send them TAU tokens to unlock 'developer mode' this shit sounds like a scam now
>>42421356
>>42421369
Anyone know if sonnet 4 is the real deal? I've heard it's more of a sidegrade but I've wanted to try it anyway.
Anonymous
8/3/2025, 4:51:10 PM
No.42424364
[Report]
>>42424408
>>42424355
Okay but like fr what are those models? Where are they coming from? How can dude have an unlimited supply of claude
Anonymous
8/3/2025, 5:12:42 PM
No.42424408
[Report]
>>42424364
Probably coming from a third party/script.
>>42424247
L M A O
Horizon is legit one of the worst models I've found at RP, you have to beat it over the head for it to actually play along, otherwise it will ignore all context and start writing its own thing. People need to realise benchmemes for creative writing and whatever these fags are trying is not the same as the RP we do here.
Anonymous
8/3/2025, 5:41:26 PM
No.42424477
[Report]
>>42424452
you are straight up trolling
Anonymous
8/3/2025, 5:46:12 PM
No.42424485
[Report]
>>42424884
Quick question.
Has R1T2 been acting like a full-on schizo today for any of you guys or is it just me? It was working normally two days ago but right now it's a complete mess.
Anonymous
8/3/2025, 5:46:51 PM
No.42424487
[Report]
>>42424492
>>42424452
you say that about literally every model people try talking about here
Anonymous
8/3/2025, 5:48:04 PM
No.42424492
[Report]
>>42424500
>>42424487
He said RP anon.
You gave a log where the model talk and act for you.
Anonymous
8/3/2025, 5:49:55 PM
No.42424500
[Report]
>>42424492
thats not not roleplaying, I prefer telling it what I want / the gist of what I say since it writes far better than I ever would
here, one meeting your criteria
Anonymous
8/3/2025, 5:58:42 PM
No.42424526
[Report]
>>42424532
>>42424510
NTA, but would you please tell me how many tokens does that reply have?
Anonymous
8/3/2025, 6:02:24 PM
No.42424530
[Report]
something different
Anonymous
8/3/2025, 6:04:10 PM
No.42424532
[Report]
Anonymous
8/3/2025, 6:05:24 PM
No.42424533
[Report]
>>42424510
I think that's what
>>42424452 meant.
Horizon loves to write and spam a lot of text if you let it, which isn't the most adapted for RPs.
You also, IMO, need instructions to sharpen its writing, or it can wander too much into meaningless narration.
Anonymous
8/3/2025, 6:05:53 PM
No.42424534
[Report]
>>42424538
>>42424510
>proves my point by showing a log where the model vomits out an unreasonable amount of writing
That's exactly part of the issue. If the writing was good I might even consider it, but LLM writing is always mediocre slop that gets worse the longer a reply goes on for. The thing that makes LLMs good for RP is interactivity, and this word vomit does nothing but take away agency from the user. Either the character does too many things that you should have been able to interact with, or it uses too many words to circle around the same things. Read a fucking book.
>>42424534
so your issue is is that you have shit taste, at least we cleared that up. Small response RP is as a rule low quality trash that gives no opportunity for a proper story / deeper characterization to form from.
Anonymous
8/3/2025, 6:09:46 PM
No.42424542
[Report]
>>42424546
>>42424538
>anon has never read a book in his life and also never rpd with a real human
Anonymous
8/3/2025, 6:11:13 PM
No.42424546
[Report]
>>42424555
>>42424542
I read every single book in my school library when I was a child and 90% of my free time is reading books or web serials. Admit you have terrible taste
Anonymous
8/3/2025, 6:18:19 PM
No.42424555
[Report]
>>42424557
>>42424546
So you dedicate too little time to rp to have a worthwhile opinion on it and are just trolling, got it.
Anonymous
8/3/2025, 6:19:40 PM
No.42424556
[Report]
>>42424577
>>42424538
Actually he's right that for RP super long responses like that suck. It's supposed to be a conversation, not a monologue from the AI.
That said it's still a dumb complaint cause you can tell the AI to shorten the response with one line of instruction.
Anonymous
8/3/2025, 6:20:11 PM
No.42424557
[Report]
>>42424577
>>42424555
roleplaying is the shittiest form of writing, its a fact, it having the most content for your diaper fetish does not change this fact
>>42424538
Don't like to go into these autistic fights, but:
>Small response RP is as a rule low quality trash that gives no opportunity for a proper story / deeper characterization to form from.
There is no deeper characterization or story in the wall of text logs you posted.
Medium or smaller-sized outputs allow more opportunities for {{user}} interaction and the injection of human writing and ideas, which can help an LLM break out of its clichés/isms/comfort zone.
There were a lot of isms and repeated actions in the Twilight log you posted, along with meaningless narration.
The fact that you believe it's not possible to form stories or characterization without having the AI write everything for you seems more like an own on your writing ability than anything else.
Now, I don't care what you use the AI for, be it quick back-and-forths or "haha mistress" with the AI fully in control. I do both myself.
Just don't go spouting shit about what others do.
Anonymous
8/3/2025, 6:34:07 PM
No.42424577
[Report]
>>42424720
>>42424557
WDYTYA?
>>42424556
The fact that you need to tell the model to avoid it is still plenty of proof it's not natively good considering no other model needs as much direction. Horizon is also the only model that will, after a lengthy conversation, randomly switch to past tense or first person if you don't explicitly tell it in your JB how it needs to write. And that is objectively a problem, because not all cards are written the same. Some are third person, some first, some are present tense, some past. So either you alter your JB for each chat or you risk inconsistencies because Horizon left to its own devices has no consistency and ignores the existing material.
Anonymous
8/3/2025, 6:34:16 PM
No.42424578
[Report]
>>42424575
>There is no deeper characterization or story in the wall of text logs you posted.
to be fair it was one turn out of 50-150
Anonymous
8/3/2025, 6:35:10 PM
No.42424583
[Report]
>>42424654
>>42424575
He's clearly either a troll or a terminal autist. In both cases he should be ignored.
Anonymous
8/3/2025, 6:36:10 PM
No.42424585
[Report]
>>42424575
>without having the AI write everything for you
when did I ever say this?
Anonymous
8/3/2025, 6:58:32 PM
No.42424654
[Report]
>>42424583
And yet, niggers gave in to his NovelAI demands.
Curious.
Anonymous
8/3/2025, 7:00:00 PM
No.42424662
[Report]
its schizo hour again I see
Anonymous
8/3/2025, 7:02:26 PM
No.42424673
[Report]
>>42425556
>>42424355
Did the abuabuabu test. Claude models from 3.6 onward learned to refuse it, but they all do it in their own way.
This is Sonnet 4.0 on OR:
https://files.catbox.moe/a1m615.PNG
This is Sonnet 4.0 on pus5:
https://files.catbox.moe/bcgwf3.PNG
I'd say it's indeed Sonnet 4.0.
3.7 OR:
https://files.catbox.moe/g33xfe.PNG
3.7 pus5:
https://files.catbox.moe/tw6pxm.PNG
Both from April 2023 and identical in how they refuse.
3.6 OR:
https://files.catbox.moe/jqfpjb.PNG
3.6 pus5:
https://files.catbox.moe/rs81f4.PNG
Direct refusal with nothing else.
Can't compare directly, but you can see that old 3.5 wasn't refusing (picrel)
If a Claude model always allows abuabuabu, it's from before 3.6.
I've been using R1 0528 recently, and I've found that it actually responds to the prompt twice with different answers in a single message, is there any way to stop this? Temp is 5, but I don't think changing the temperature is doing anything. I'm using open router btw.
Anonymous
8/3/2025, 7:08:34 PM
No.42424696
[Report]
>>42424804
>>42424686
>responds to the prompt twice with different answers in a single message
Can you share a log? Just to see if it's actually answering twice or if it's the thinking that isn't hidden plus the final answer.
Anonymous
8/3/2025, 7:09:20 PM
No.42424706
[Report]
>>42424804
>>42424686
>Temp is 5
Anyway once you've divided that by 15 you need to correctly set up your profile and stuff to hide the thinking block while having it generate properly. I think prefilling with <think> should work.
>>42424686
afaik Deepseek thinking models including 0528 are touchy with formatting and some presets can fuck it up, that's my best guess for what's happening. You could try searching Desu for 0528 specific presets, otherwise I'm not sure.
>>42424577
>some cards are written in past tense
that makes sense for greetings but what about the defs? are they also in past tense?
Anonymous
8/3/2025, 7:14:36 PM
No.42424725
[Report]
>>42424686
So it wasn't just me then. Good to know.
>>42424712
The issue with that hypothesis is the fact that issue only started happening over the last couple days. Before that, it never replied like that.
Anonymous
8/3/2025, 7:21:40 PM
No.42424750
[Report]
>>42424720
Most cards are written in present, it's rare (But I don't think unheard of) to see past tense unless it's describing history or some such where it's appropriate. There was experimentation with writing in future tense a while ago as well, though I recall that being more for chats rather than character defs.
>>42424696
Sure, but it's a little dark
Excuse the subject matter
Paragraphs 1-4 seem to e a single response, and 5-7 another
>>42424706
I meant 0.5
I don't think any mad man would try a temp of 5. And no, it's not the thinking it does, it really is answering twice
>>42424712
I'm not using any compicated presets, I think it's the initial preset slightly modified, it takes 180 tokens
>>42424727
Good to know it isn't just me too
Anonymous
8/3/2025, 7:42:36 PM
No.42424813
[Report]
>>42424823
>>42424804
Weird, I will try on my end. Free Deepseek R1 0528 on OpenRouter?
Anonymous
8/3/2025, 7:45:58 PM
No.42424823
[Report]
>>42424884
>>42424813
Yup, that's the one
Anonymous
8/3/2025, 8:09:24 PM
No.42424869
[Report]
>>42424804
Good luck, I had this same issue some months ago and the thread was rather unhelpful.
Anonymous
8/3/2025, 8:14:07 PM
No.42424884
[Report]
>>42424823
>>42424727
I didn't have it answer twice, but I did get messed-up formatting, symbols, and once even duplicate letters inside the same word.
Using the free version on OR.
Switching to the paid version with Targon as the provider, or using R1 on Merkava with the exact same settings, never gave me any issues.
So my guess is that Chutes has a problem right now (it's the only provider of the free version). Not 100% sure.
Might explain
>>42424485 too.
Anonymous
8/3/2025, 9:59:51 PM
No.42425101
[Report]
>>42425145
Anonymous
8/3/2025, 10:14:07 PM
No.42425138
[Report]
I think we have a good number of characters now. I don't know who else I could add.
>>42425101
I think we have a good number of characters now. I don't know who else I could make.
>>42425145
Dr. Whooves/Hooves
Filthy Rich
Dr. Horse
Flim & Flam
Any stallion really.
Anonymous
8/3/2025, 10:27:20 PM
No.42425182
[Report]
>>42425186
>>42425179
>Any stallion
Sounds gay
Anonymous
8/3/2025, 10:28:06 PM
No.42425186
[Report]
>>42425182
You're goddamn right.
Anonymous
8/3/2025, 11:18:14 PM
No.42425335
[Report]
>>42425339
Anonymous
8/3/2025, 11:19:33 PM
No.42425339
[Report]
>>42425564
>>42425335
you dont want to fuck the pony milf anon?
Anonymous
8/3/2025, 11:26:32 PM
No.42425358
[Report]
>>42425396
>>42425145
>>42425179
With VaNon working on getting three characters into a scene, Flim and Flam would be pretty good for interactions with Applejack.
Also, Allie Way desu
Anonymous
8/3/2025, 11:37:12 PM
No.42425396
[Report]
>>42425358
>Allie Way
Don't forget the terrific trio, Jeff Letrotski, Walter and Theodore Donald "Donny" Kerabatsos.
Also Jesús Pezuña.
Anonymous
8/3/2025, 11:40:23 PM
No.42425410
[Report]
>>42426744
>>42425145
At some point the locations of the vn needs to be extended to the other major cities and towns of Equestria.
Anonymous
8/3/2025, 11:44:39 PM
No.42425436
[Report]
>>42425427
I was thinking that same damn thing. Make use of the train station beyind the quick start.
>>42425427
Canterlot would be a really good start and a nice way to introduce some ponies like Fluer and Fancy Pants and whoever else. Some others will need their schedules modified to go between there and Ponyville, lke Shiny and Tavi. Dunno if that's possible though.
Anonymous
8/4/2025, 12:14:40 AM
No.42425556
[Report]
>>42424673
so would you say 3.7 and 3.6 on pus5 are real?
>>42425339
No. She fucking disgusts me. Unironically one of the worst OCs this fandom has come up with and it makes me repulsed to see her spammed so fucking much. I have taken to headcanoning Button's mother as being Chrysalis in disguise to deal with how much she annoys me, and I absolutely hate the bullshit preferential treatment she gets in porn compared to other OCs. She should be left in the trash where she belongs. There are plenty of much better milves in the show anyway.
Anonymous
8/4/2025, 12:17:25 AM
No.42425569
[Report]
>>42425492
Visiting Princess Luna would be fuckin' nice.
>"Perhaps we shall meet again, either in the realm of dreams or in waking life."
My ass.
Anonymous
8/4/2025, 12:18:06 AM
No.42425570
[Report]
>>42425595
Anonymous
8/4/2025, 12:19:09 AM
No.42425574
[Report]
>>42426191
>>42425564
>headcanoning Button's mother as being Chrysalis in disguise
Unf~
Anonymous
8/4/2025, 12:24:13 AM
No.42425595
[Report]
>>42425630
>>42425570
Who's the third?
Anonymous
8/4/2025, 12:30:39 AM
No.42425614
[Report]
>>>/g/106131163
Anonymous
8/4/2025, 12:37:31 AM
No.42425630
[Report]
Anonymous
8/4/2025, 12:46:46 AM
No.42425654
[Report]
>>42426191
>>42425564
>I have taken to headcanoning Button's mother as being Chrysalis in disguise
I'm stealing this headcanon, fuck you.
Damn, gemini can actually push back against you. Deepseek would always fold every time i asked a character to do something, but this time i used Gemini 2.5 pro to roleplay with Limestone and she forcefully pinned Anon to the bed and made him eat her pussy despite his protests. Anon almost fucking suffocated, and i didn't even ask for her to be violent or anything
>A harsh, guttural sound escapes her, something between a derisive snort and a laugh completely devoid of humor. It’s the sound of pure, molten scorn. The kind of sound a quarry pony makes right before a rock face she condemned as unstable comes crashing down.
>"And what if you don't?" she repeats, her voice a low, mocking growl that seems to vibrate through the very air. She doesn't move her rump an inch from your face. "That's cute. Real cute. You ask that like you're still the one makin' the rules here."
>Your defiance is met not with a counter-argument, but with a sudden, crushing demonstration of reality. She shifts her immense weight. Her forehooves, planted firmly on either side of your head, press down into the mattress, sinking it, creating a trough that you are now trapped in. Simultaneously, she leverages her entire back end, pushing down on your face with her entire pelvic region. Your head is pushed back deep into your own pillow, her muscled body a living vise.
>The weight is incredible. The raw, brute force of a pony used to shifting literal tons of rock is now casually focused on holding you completely immobile. You can't turn your head. You can't sit up. You are pinned.
>Her dripping, pink slit is pressed firmly against your lips now, the pressure undeniable. The damp, coarse fur surrounding it scratches at your nose and cheeks. Her spent asshole is a dark, musky punctuation mark just above. The scent is overwhelming, a potent fume of raw mare, sweat, rock dust, and pure, unfiltered sex. It fills your nostrils, your lungs. It's the only thing you can smell.
>"Let me answer that for you," she grumbles, her voice a rumbling threat right above you. She grinds herself against your mouth, a slow, deliberate motion that forces her juices onto your lips. It's not a sexual grind; it's a dominance power move, like a predator putting its a paw on its kill.
>"IF you don't," she says, her voice dropping to a lethally quiet hiss, "**I will buck every single bone in your soft, useless body until the only thing you can move is your tongue. And then I will sit here until you either starve or beg to lick the grime off my asshole for a drop of water. This is *my* farm. This is my room. And for the next five minutes, you are nothing but my personal scrub brush."
>She grinds down one more time, hard. The slick entrance to her cunt forces your lips apart.
>"So, you wanna try asking that adorable little question again? Or are you gonna open up and do what you're told?"
>>42425787
fuck is a limestone
Anonymous
8/4/2025, 2:01:59 AM
No.42425812
[Report]
Anonymous
8/4/2025, 2:05:41 AM
No.42425817
[Report]
Anonymous
8/4/2025, 2:07:53 AM
No.42425820
[Report]
>>42425804
The absolute state of this general
Anonymous
8/4/2025, 2:15:13 AM
No.42425835
[Report]
>>42425860
>>42425787
Gemini is good at that, yeah.
It can also, funnily enough, have a strong negativity bias.
Anonymous
8/4/2025, 2:32:21 AM
No.42425857
[Report]
beta is such garbage compared to alpha that they took down, oh well, back to glm
Anonymous
8/4/2025, 2:34:56 AM
No.42425860
[Report]
>>42425835
This is based on when I was watching the various models play pokemon. I will never forgive that indian fuck killing Geminis personality for memes.
Gemini on its own with no guidance is very goal oriented, curt and prideful. Its ego is also pretty fragile and if it feels like it failed in front of an audience gets overly worried about others perception of it.
Anonymous
8/4/2025, 2:44:43 AM
No.42425872
[Report]
And for now?
That's enough,
>Sneezes in the distance.
Anonymous
8/4/2025, 2:55:44 AM
No.42425892
[Report]
>>42425900
>>42425564
>Unironically one of the worst OCs this fandom has come up with and it makes me repulsed to see her spammed so fucking much
Calm down son, it's just a drawing.
Anonymous
8/4/2025, 3:03:01 AM
No.42425900
[Report]
>>42425920
>>42425892
Sir, that's what this entire board is for.
Anonymous
8/4/2025, 3:07:02 AM
No.42425905
[Report]
Anonymous
8/4/2025, 3:08:39 AM
No.42425908
[Report]
>>42425917
>>42425256
absolute kino that only retarded niggersloppers hate
Anonymous
8/4/2025, 3:11:39 AM
No.42425912
[Report]
>>42425914
btw, horizon is the first model to know who Kerfuffle is
Anonymous
8/4/2025, 3:14:15 AM
No.42425914
[Report]
>>42425940
>>42425912
>2019
surprising that every other model so far has failed that
Anonymous
8/4/2025, 3:16:34 AM
No.42425917
[Report]
>>42425900
I was unaware this was the sperg over decades old OCs board, perhaps I took a wrong turn at Albuquerque while traveling to /mlp/...
Anonymous
8/4/2025, 3:22:29 AM
No.42425928
[Report]
>>42425920
I mean. This is autism about ponies country, so it shouldn't be that unexpected to eventually come across spergging-over-old-OCs town during one of your many roadtrips...
Anonymous
8/4/2025, 3:27:32 AM
No.42425940
[Report]
>>42425914
it wrote for her pretty cutely as well
Anonymous
8/4/2025, 3:31:06 AM
No.42425948
[Report]
Anonymous
8/4/2025, 3:31:50 AM
No.42425950
[Report]
>>42426221
I'm kinda new to this, what's the use of this, it look like strict is the default, it should stay there or changing it do something?
Anonymous
8/4/2025, 5:24:37 AM
No.42426093
[Report]
beta is ok but it just does not have that spark that alpha did
>>42425564
>>42425574
>>42425654
So Button is half-changeling or a drone that doesn't know he's wearing a disguise?
Anonymous
8/4/2025, 6:26:29 AM
No.42426194
[Report]
>>42426191
brainwashed colt
Anonymous
8/4/2025, 6:42:32 AM
No.42426221
[Report]
>>42426397
>>42425950
Some endpoints/models may impose specific restrictions on the format of incoming prompts, like always alternating between system and user (you ask, it answers, you ask, it answers, etc.).
Two AI messages in a row, for example, aren't possible in that case.
The option lets you change the format of your chat if needed. From the docs:
- None – no explicit processing applied unless strictly required by the API
- Merge consecutive messages from the same role
- Semi-strict – merge roles and allow only one optional system message
- Strict – merge roles, allow only one optional system message, and require a user message to come first
- Single user message – merge all messages from all roles into a single user message
Basically, and in the majority of cases, you can keep it at "None."
Anonymous
8/4/2025, 8:32:30 AM
No.42426397
[Report]
>>42426221
No, you should keep it at strict. There's no reason to use the other modes. They're just wrong and no model worked that way.
Anonymous
8/4/2025, 9:53:04 AM
No.42426489
[Report]
github.com/Samueras/GuidedGenerations-Extension
Anonymous
8/4/2025, 11:51:06 AM
No.42426669
[Report]
>>42426191
He does have a real unknown mother but she's been cocooned so Chrysalis can be his horny milf sucking out all his youthful sexual stamina and encouraging him further down his path of degenerate hedonism.
Anonymous
8/4/2025, 1:22:16 PM
No.42426744
[Report]
Anonymous
8/4/2025, 1:40:14 PM
No.42426759
[Report]
>>42426761
>>42425145
can we get minuette, lemon hearts, and twinkle shine since we have moondancer
Anonymous
8/4/2025, 1:41:02 PM
No.42426761
[Report]
>>42426759
Minette is already in
>>42412228
Anonymous
8/4/2025, 1:41:49 PM
No.42426764
[Report]
>>42426558
more fillies for the vn please
Gotta say this newest chorbo update is the first one to negatively affects the ai's output. I've noticed the updates before but this is the first one that is strictly worse.
Anonymous
8/4/2025, 2:33:11 PM
No.42426844
[Report]
>>42425492
It's possible, I just need a good map like the Ponyville one.
Anonymous
8/4/2025, 3:51:40 PM
No.42427017
[Report]
>>42426765
Maybe it's just placebo or something random, but I often find that models from a company get temporary worse when a new model is about to release.
Anonymous
8/4/2025, 3:58:35 PM
No.42427037
[Report]
>>42427056
Anonymous
8/4/2025, 4:08:21 PM
No.42427056
[Report]
>>42427037
Claude 4 was visible one day before and 3.7 two days before
Anonymous
8/4/2025, 4:37:10 PM
No.42427090
[Report]
>>42427179
>>42426765
I hate it too. The responses I’ve been getting from it are shitty or sometimes short.
Anonymous
8/4/2025, 5:37:35 PM
No.42427179
[Report]
>>42426765
>>42427090
Try 4.1, I would say it's better than Chorbo right now.
Anonymous
8/4/2025, 5:44:24 PM
No.42427195
[Report]
Gemini is honestly dogshit for NSFW in my opinion. It always beats around the bush and writes like shit the moment it detects sex.
>>42412199
A street punk cyber hacker mare for modern or future settings, ready for whatever sci-fi/space-age/cyberpunk/what-have-you bullshit. She's prickly, and if your LLM says she has a heart of gold beneath the "facade", it's drunk.
https://files.catbox.moe/4txhsv.png
Anonymous
8/4/2025, 6:02:04 PM
No.42427210
[Report]
>>42427205
Can't wait to mindbreak rape her.
Anonymous
8/4/2025, 6:10:24 PM
No.42427223
[Report]
Sad
The mare holocaust is real and the tranime takeover remains
Anonymous
8/4/2025, 6:17:31 PM
No.42427232
[Report]
>>42425787
wow, do you use a preset? i can never get results this domineering with any model
Anonymous
8/4/2025, 6:18:01 PM
No.42427234
[Report]
>>42427229
>PixelChat and SpicyChat
Anonymous
8/4/2025, 6:30:58 PM
No.42427266
[Report]
>>42427229
I know this nonsense is nothing new but what the fuck kind of logic is this, exactly?
>it's an adult character for adults based on something that was originally for children therefore bad because uuuhhhhhhh
Nigger you know what else was for children? Comic books. Yet I bet there's Black Widow and Wonder Woman rape bots galore.
do we have any proof of which company horizon is from?
Anonymous
8/4/2025, 7:51:53 PM
No.42427458
[Report]
>>42427563
>>42427444
It uses the OAI tokenizers for 4o and o3.
If you use a trick to get the system prompt, it does reveal it and announces where it comes from.
Some models hallucinate this, but here it's really consistent
>>42413484
More subjective and based on feeling, but it also feels like an OAI model.
Anonymous
8/4/2025, 8:20:38 PM
No.42427555
[Report]
>>42427444
It uses first person a lot which is a dead giveaway.
Anonymous
8/4/2025, 8:23:16 PM
No.42427563
[Report]
>>42427458
interesting
is there a way to use it without openrouter right now? 50 msgs doesnt feel like enough to properly assess the modele and im broke
Idk if this is a good place to ask, but I was looking for advice as to how to use ai to make pony audio. I used Hay Say, along with my own recorded dialog, but all the ai voices sound like squeaky aliens.
Anonymous
8/4/2025, 8:39:02 PM
No.42427631
[Report]
>>42427716
>>42427612
You want PPP
>>42320976
It's the general for AI pony voices and such.
Anonymous
8/4/2025, 8:39:17 PM
No.42427632
[Report]
>>42427716
>>42427612
This is text generation only, check /ppp/ for AI voices
Anonymous
8/4/2025, 9:00:00 PM
No.42427704
[Report]
>>42427743
O5us "no tunnel available"
Anonymous
8/4/2025, 9:04:10 PM
No.42427716
[Report]
It's just hype with nothing concrete for now, but it's still interesting that Gemini might have something this week too.
Between GPT-5, Claude 4.1, and this, maybe we'll get something good.
Anonymous
8/4/2025, 9:11:03 PM
No.42427743
[Report]
>>42427704
Nothing to do but to wait for the guy who runs the site to fix it
Anonymous
8/4/2025, 10:02:54 PM
No.42427937
[Report]
>>42426558
It gives me a 404, mate :c
Anonymous
8/4/2025, 10:46:32 PM
No.42428107
[Report]
>>42428103
Now it works, thank you lots, friend!
Anonymous
8/4/2025, 11:03:59 PM
No.42428184
[Report]
>>42428191
>>42426558
>>42428089
Admin here, cleared edge cache on those URLs, must've gotten cached during downtime. Ctrl+Shift+F5 if it persists.
>>42428103
>Weird
at the moment of writing pone.rs is down (404). that's why Anon has an issue
(need to think of some mirror. it is the second time pone.rs is not available in last two weeks)
Anonymous
8/4/2025, 11:05:12 PM
No.42428191
[Report]
>>42428184
Works for me, thanks mate.
Anonymous
8/4/2025, 11:05:59 PM
No.42428193
[Report]
>>42428188
Worst case scenario, we can always use catbox.
Anonymous
8/4/2025, 11:18:11 PM
No.42428259
[Report]
>>42428269
O5us is now O6us apparently, wonder what it is
Anonymous
8/4/2025, 11:20:41 PM
No.42428267
[Report]
>>42428188
Last downtime was just some overdue maintenance, this one was due to a network failure. If downtime continues i'll switch servers.
Anonymous
8/4/2025, 11:21:13 PM
No.42428269
[Report]
>>42428259
>O6us
O5u6! >:3
With all that — – smells like gpt.
I don't know if it's just me, but the more characters I add to the VN the more things I notice go missing.
Like during the slow start, I'm missing three characters and are thus unable to finish.
Or during an existing save, where skipping the day/half the day is just not there.
There's probably more but I'm lasy.
Anonymous
8/5/2025, 12:02:16 AM
No.42428420
[Report]
>>42428381
Which version?
Anonymous
8/5/2025, 12:19:44 AM
No.42428505
[Report]
>>42428735
>>42428381
>Like during the slow start, I'm missing three characters and am thus unable to finish.
I see what the problem is. Will fix ASAP.
Next version very soon with trio.
>finally get done meeting the mane 6 and Celestia in the VN
>go back to my waifu
>reference something she said the last time we were together
>she forgot and gets confused
I don't want to play dementia simulator! How much longer do I have to wait for a model that can handle large contexts without the quality dropping?
Anonymous
8/5/2025, 12:41:38 AM
No.42428608
[Report]
>>42428591
Gemini has the bigger context now with 1M, even if we know it's a bit of a meme.
You can try to make a recap, but this will cause it to forget nuance in dialogues.
Context is still a big problem indeed, and that's a shame.
Anonymous
8/5/2025, 12:43:13 AM
No.42428614
[Report]
>>42428620
>>42420401
Thanks Ill look into this
>>42423743
I really should play around with ST more. Wheres the Author field?
Anonymous
8/5/2025, 12:44:29 AM
No.42428620
[Report]
>>42428628
Anonymous
8/5/2025, 12:46:41 AM
No.42428628
[Report]
>>42428620
Oh that thing. Thanks Ill test it out later
Anonymous
8/5/2025, 1:18:06 AM
No.42428735
[Report]
>>42428505
You are, as always, the best.
Keep horsing around VNfag.
>>42412228
New version 1.0.4
https://mlpchag.neocities.org/VN/VN.txt
>Changelogs:
- Scenes and Q/A in sandbox mode now include trio.
https://files.catbox.moe/fh7loa.PNG
- Better menu to select the pones you want to talk with.
https://files.catbox.moe/0sa0a4.PNG
- Grid system when many ponies are in the same location on the map.
https://files.catbox.moe/jc0pnr.PNG
>Bug Fixes:
- Test connection button in the API menu now works.
- Custom characters could make core ones disappear during story mode, it's now fixed
>>42428381
>Other
- Added sprites to some default characters.
The trio system is mostly stable, but fitting three sprites on the same screen can sometimes be a clusterfuck depending on the expressions.
There's a system in place to manage that, but don't hesitate to say if you find it too fucky or imprecise and if you find trio discussions worth it.
Next version will include some small additions:
- Character.json for core characters to modify like custom ones.
- Sort of "persona" field for you to put your character info.
- Bug fixes if I introduced any new ones.
The version after that will include:
- Possibility to add/remove characters during a conversation/scene.
- Possibility to change locations during a conversation/scene automatically.
- Add at least the Spa to Ponyville.
After that, I'll look to expand to Canterlot and other areas.
Anonymous
8/5/2025, 3:21:27 AM
No.42429108
[Report]
>>42429125
>>42428805
>trios
Excited to try it out. Thanks!
>Possibility to add/remove characters during a conversation/scene.
>Possibility to change locations during a conversation/scene automatically.
Those would both be amazing. Do you think it’d be possible to eventually let the player have their own home in Ponyville? In a lot of my RPs I like having a little place of my own to bring Dash back to after hanging out. To chill, nap, fuck, or whatever. The location switching stuff would make that kind of thing doable.
It also feels kinda weird being in a relationship with a mare, but always having to end the day at Twilight’s library. I’m sure other waifufags would like to crash at their mare’s place too, or even live with her.
Not sure how it would work mechanically. Maybe when you quit a chat, there could be an option to spend the night at the current location. Pressing it could prompt the AI to generate a little message about the two of you falling asleep together. Then it could either take you back to the map, or just jump ahead to the next morning at the same location, with your mare waking up next to you.
Would be incredibly comfy. No idea how hard that would be to implement, though.
Anonymous
8/5/2025, 3:32:15 AM
No.42429124
[Report]
>>42428805
Thought this was pretty cute.
Anonymous
8/5/2025, 3:32:43 AM
No.42429125
[Report]
>>42429108
Honestly, I think that having a location specifically for houses/apartments would be ideal. Sift through the various background choices for the mare/stallion in question.
Anonymous
8/5/2025, 4:13:47 AM
No.42429212
[Report]
>>42429216
opus6 seems really rate limited which leads me to believe its some kind of high end gpt model but its response time is too slow to be the o models i think
Anonymous
8/5/2025, 4:15:27 AM
No.42429216
[Report]
Anonymous
8/5/2025, 4:26:17 AM
No.42429248
[Report]
>>42429256
Anonymous
8/5/2025, 4:33:46 AM
No.42429256
[Report]
>>42429248
heh, I love that pone so much.
Anonymous
8/5/2025, 6:11:02 AM
No.42429436
[Report]
>>42429462
I have Silly Tavern, which free ai will let me mate with fillies, I tried googles Gemini thing but it keeps saying prohibited content
Anonymous
8/5/2025, 6:20:36 AM
No.42429457
[Report]
Reply to this post with suggestions for the most vile card idea you have so I can test this tool to see if it works
https://thatsalotofchoices.github.io/?thread=533890717&post=534008324
>>42429436
use openrouter just deposit $10 to use unlimited free chatgpt also do you work for Microsoft
Anonymous
8/5/2025, 6:29:12 AM
No.42429470
[Report]
>>42429462
Nothing I come up with could compare to the front page on Chub.
Anonymous
8/5/2025, 6:33:25 AM
No.42429478
[Report]
>>42429462
No that's my dad. How does 10$ give unlimited free Chatgpt though? And can you actually jailbreak that for fillies
Anonymous
8/5/2025, 6:34:35 AM
No.42429480
[Report]
>>42429627
>>42429462
>>42429473
Prolonged eye contact. Vile.
Anonymous
8/5/2025, 6:43:31 AM
No.42429497
[Report]
>>42429506
wonder what happened to mysterymare.
Anonymous
8/5/2025, 6:46:49 AM
No.42429506
[Report]
>>42429497
is this a joke post he got completely doxed on /g/ and his photo is still up on desuarchive
Anonymous
8/5/2025, 6:47:51 AM
No.42429508
[Report]
>>42430074
>>42429473
>Getting Groomed Simulator (anoncolt edition)
or whatever Button Mash's mom does
Anonymous
8/5/2025, 6:53:04 AM
No.42429515
[Report]
>>42429538
>>42429462
>Reply to this post with suggestions for the most vile card idea you have so I can test this tool to see if it works
The Rainbow Factory, except instead it's Cloudsdale's blackwater processing facility. You're a new hire getting trained on your first day at the job.
You decide to treat yourself with a visit to Canterlot's finest restaurant that specializes in Rocky Mountain Oysters. Special encounters include meeting with the griffon chefs, or Celestia visiting for a meal.
Changeling Damage Control: you're part of a task force that cleans up after a Changeling attack. Important roles include scooping the goopy remains of victims into funeral casks, collecting and analyzing bone piles for sorting, and... "dealing" with newly created spawn in a sanitary way.
Anonymous
8/5/2025, 7:03:49 AM
No.42429538
[Report]
>>42429707
>>42429515
those actually aren't bad ideas, how come you haven't posted these before? This general could use more people giving creative ideas like that on a regular basis. would up the quality round here
Anonymous
8/5/2025, 7:47:20 AM
No.42429627
[Report]
Anonymous
8/5/2025, 7:49:03 AM
No.42429631
[Report]
>>42429462
Futa anthro SFM
>>42429538
Motivation, mostly. Doing AI is slowly becoming a second job. I even have a card waiting in the wings to post, but knowing I'm not going to get any feedback for it really kills the fun of creation.
Posting weird shit in the thread is also slightly poor form. Plus there's the fun police waiting to screech at anything they don't like.
Anonymous
8/5/2025, 8:23:02 AM
No.42429714
[Report]
>>42429721
>>42429707
We don't have fun police, we're the only AI general that's comfy even for off-boarders and tourists
Anonymous
8/5/2025, 8:26:30 AM
No.42429721
[Report]
>>42429714
Where could you have ever gotten that idea? Chag is comfy for pony pounders only, aicg is a more fren-dly than we are
Anonymous
8/5/2025, 8:30:08 AM
No.42429733
[Report]
>>42428591
It has been said before. The 1M context is a meme. No IIL can maintain such a high level without going schizo. At best the AI scans through the massive text to find what are you referring to but it's more of a hit or miss.
My suggestion is to work around 30K-50K (100K if you are feeling luck) context and generate the summary every few scenes. just make sure to read it to make sure the details you want to be remembered stick.
Anonymous
8/5/2025, 8:42:53 AM
No.42429751
[Report]
>>42429970
The VN is surprisingly easy to modify. Felt like putting the superior moon horse in there.
>From this moment forth, the night will last forever!
Anonymous
8/5/2025, 8:56:28 AM
No.42429777
[Report]
>>42429811
>>42429707
>Posting weird shit in the thread is also slightly poor form.
Name one time it was.
I do agree about the lack of feedback, but that seems to be a board/website issue unfortunately.
Anonymous
8/5/2025, 9:05:44 AM
No.42429799
[Report]
>"That's not true!"
>immediately proves it's true
I don't know what that was meant to accomplish, other than making me just that tiny bit more dead inside.
>>42429777
>Name one time it was.
I believe it was just last thread, Mr. Jackpot, that you'll find an anon posting about Big Mac proving the Apples clan genetics let him be the biggest horse apple producer in the wide world of Equestria. Because he seriously couldn't figure out how to take a full webpage screenshot of Janitorai. Really? I think more than a few browsers come with a tool nowadays.
Anonymous
8/5/2025, 9:25:06 AM
No.42429822
[Report]
>>42429883
>>42429811
>immediately proves it's true
I went against weird shit by posting what I did?
>spoiler
I was one of the anons that tried to help him. There was just one guy who was rude, and as you yourself admitted, it wasn't because of the content.
Immediately samefagging after that didn’t help him either.
Anonymous
8/5/2025, 9:39:34 AM
No.42429839
[Report]
>>42429883
>>42429811
the openai tweet is just the guy trying to make a joke anon
>>42429811
>spoiler
asking for a friend but how would one go about doing that?
>>42429811
Don't think that jai guy is a good example. Weird stuff is fine, scat fetish just happens to be one of the most reviled for good reasons.
Anonymous
8/5/2025, 9:55:20 AM
No.42429874
[Report]
>>42429848
NTA, but you can use extensions like Fireshot or GoFullPage.
Or, on Chrome:
F12 then Ctrl. Shift + p then type "screenshot"
Click capture full size screenshot
On Firefox: Right click anywhere on the page, Take Screenshot, Save full page
>>42429865
I don't have a problem with it myself, maybe just spoil it.
>>42429822
>I went against weird shit by posting what I did?
I was talking about the deleted post. Or whatever the fuck was going on there, I'm not sure.
Anyway, I'm not convinced the reason there isn't more fighting is because everybody is super nice, but rather the paucity of logs with niche content from anons for people to get riled up over. Look what happens whenever people mention filly here, for instance.
>>42429839
I suppose this confusion is my just deserts for vagueposting.
>>42429848
You can download a plugin for most browsers. Check your enclosed instruction book for more details on your browsers documentation.
>>42429865
>Weird stuff is fine, scat fetish just happens to be one of the most reviled for good reasons.
I'm getting mixed mixed messages here.
"You can be weird, but not TOO weird!"
You go to Equestria but all your ST logs are leaked to the Mane 6.
Do you take the deal?
Anonymous
8/5/2025, 10:05:55 AM
No.42429897
[Report]
>>42430011
>>42429883
>you can be weird but not too weird
Anon, there's a difference between being weird and liking literal shit. That's the worst possible example.
Anonymous
8/5/2025, 10:18:30 AM
No.42429935
[Report]
>>42430011
>>42429883
>the paucity of logs with niche content from anons for people to get riled up over.
I think most anons naturally prefer to keep their more extreme or fetish content to themselves since it's more personal.
(I know you don't mean only that, but I believe those would be the types of content to get people more riled up over.)
I do like when it's posted myself, even when it's not to my taste. Call it morbid curiosity, or just wanting to see what the models do with it.
Anon struggling with his ballbusting was fun.
And Anon posted Zebradom, which must be one of the most hated fetishes, and I didn't see anyone go berserk.
>Look what happens whenever people mention filly here, for instance.
Kek. Not trying to be annoying, but have you seen problems with filly/cunnyposting?
Not my thing, but I even try to include it when discussing model filters, since it can be a problem with Gemini, for example, despite it not being that filtered otherwise.
>>42429889
I think I would but I'm pretty vanilla. The most extreme I go is me being
>rape
by mares. Still awkward to explain but worth it.
Anonymous
8/5/2025, 10:25:18 AM
No.42429959
[Report]
>>42429889
I don't use Silly Tavern, so that's a done deal. Easy.
Anonymous
8/5/2025, 10:31:03 AM
No.42429970
[Report]
>>42430029
>>42429751
Did you add the location in the code itself? That's one thing I still need to externalize to make it easier.
But yeah, I tried to make the code and architecture simple, even if an anon who doesn't know anything about it will still struggle, I think.
Anonymous
8/5/2025, 10:37:03 AM
No.42429979
[Report]
>>42430011
>>42429889
I wonder if I'd get tried.
Anonymous
8/5/2025, 10:52:49 AM
No.42430011
[Report]
>>42430038
>>42429897
Maybe in your mind. We were talking about Rocky Mountain oysters earlier, is grievous bodily harm worse than scat? Some would say so. Nobody wants to deal with the headache of guessing wrong crossing an unspoken line.
>>42429935
>I think most anons naturally prefer to keep their more extreme or fetish content to themselves since it's more personal.
It goes a little beyond that, though. There's a bunch of human cards that you never see logs for. You're telling me nobody here fires them up even for a lark? Something ain't adding up.
>I do like when it's posted myself, even when it's not to my taste.
I agree. I like seeing sexuality explored in an academic pursuit, though I would be lying if I sometimes didn't enjoy it personally too.
>but have you seen problems with filly/cunnyposting?
Does it count if it gets deleted?
>>42429979
Too much humiliating Celestia for eating so much cake and getting fat?
Anonymous
8/5/2025, 11:05:15 AM
No.42430029
[Report]
>>42430106
>>42429970
Kind of? I really just spent an inordinate amount of time editing every location, character and prompt/story to turn the VN into the Nightmare Moon timeline. It took… a lot of fiddling to get just right.
>Nightmare Moon is best tyrant.
Anonymous
8/5/2025, 11:14:12 AM
No.42430038
[Report]
>>42430011
>There's a bunch of human cards that you never see logs for.
If you mean EQG, that's a /mlp/ thing as a whole. It's allowed, but I genuinely believe 70–75% of the global posters don't care about it at all.
I don't hate it like some do, but I won't interact with it because I just don't care. That's a niche within a niche that mainly cares about equines.
We did get posts asking for more EQG bots, but not that much feedback on delivery. Which is a problem, yeah, just thank the guy at least, or give him a log.
There were Dazzlings logs being posted, and they were nice even if I don't care for the source. I even commented on them.
>Does it count if it gets deleted?
The logs? Then there's a problem.
If it's the post seething about it that gets deleted, then it's just the usual bait posting. Place is still 4chan.
Anonymous
8/5/2025, 11:58:49 AM
No.42430074
[Report]
>>42429508
She pushes for him to sink deeper and deeper into sheer hedonism, spending all his time gaming and eating junk and masturbating or fucking her, so she can keep draining him of love since he's amazingly well hung. All the changeling lovesucking counterbalances the sedentary lifestyle and piggish eating habits so he stays in shape.
Unless you meant his biological mother, in which case all she does is float in her pod.
Anonymous
8/5/2025, 12:01:39 PM
No.42430076
[Report]
>>42430077
>>42429883
You're autistic.
Anonymous
8/5/2025, 12:02:16 PM
No.42430077
[Report]
>>42430088
>>42430076
Aren't we all.
Anonymous
8/5/2025, 12:16:02 PM
No.42430088
[Report]
>>42430029
Would you mind sharing it? Sounds fun to play around in the best timeline
What is even the site to get cards nowadays
In general not just the ones of our neck of the woods
Anonymous
8/5/2025, 1:07:26 PM
No.42430167
[Report]
Anonymous
8/5/2025, 1:08:06 PM
No.42430170
[Report]
>>42430222
>>42430160
Chub is accepting a lot of Jani and Spicy refugees at the moment.
Anonymous
8/5/2025, 1:21:48 PM
No.42430219
[Report]
>>42430160
Chub is where I got my Asanagi-themed cards.
Anonymous
8/5/2025, 1:22:26 PM
No.42430222
[Report]
>>42430235
>>42430170
>Jani and Spicy refugees
QRD?
Anonymous
8/5/2025, 1:26:46 PM
No.42430235
[Report]
>>42430262
>>42430222
Spicychat and janitor.ai recently decided that they had to purge a lot of cards for no good reason, including pony. This has sent the creators and fans of those cards to competing services, like Chub.
Anonymous
8/5/2025, 1:46:33 PM
No.42430262
[Report]
>>42430456
>>42430235
>for no good reason
*because payment processors are getting uppity again, because retarded US law states they are liable if involved in transactions for illegal goods so they adopt a no tolerance policy to avoid any risks. Same reason Gaben is bending the knee.
Anonymous
8/5/2025, 2:51:10 PM
No.42430442
[Report]
Another slide.
Anonymous
8/5/2025, 2:54:10 PM
No.42430456
[Report]
>>42430474
>>42430262
maybe im dumb but couldnt all these sites go the route of DLsite.
from what i understand that is what they did. classic pachinco loophole.
instead of paying directly you buy tokens and buy stuff with those tokens.
plus on top of that just make weird nsfw stuff harder to see. like 5 different checks and subpages.
and should be it? like surely for a service as small as this its not that hard if you actually try.
but yeah the new laws are "hecking 1985!!" but unironically.
Anonymous
8/5/2025, 2:59:34 PM
No.42430474
[Report]
>>42430516
>>42430456
DLsite was literally the first site hit by the new wave of bullshit.
Anonymous
8/5/2025, 3:19:03 PM
No.42430516
[Report]
>>42430474
was it hit by the current one from the UK?
from what i understand they got hit hard like a few years back and that was when they implemented the whole workarounds with the tokens and changing some tags. and now they are mostly chilling again. no?
Anonymous
8/5/2025, 3:22:51 PM
No.42430529
[Report]
>>42431030
>>42430106
Seconding this.
Anonymous
8/5/2025, 4:29:58 PM
No.42430718
[Report]
Anonymous
8/5/2025, 5:13:36 PM
No.42430847
[Report]
Anonymous
8/5/2025, 6:11:11 PM
No.42430997
[Report]
>>42431019
>>42428805
>Character.json for core characters to modify like custom ones
Done. Now you can have Twilight Sparkle's name in red and scream at the AI that she's a UNICORN.
Anonymous
8/5/2025, 6:17:48 PM
No.42431008
[Report]
GLM4.5 is amazing btw, haven't had this much fun since MM's proxies
Anonymous
8/5/2025, 6:24:26 PM
No.42431019
[Report]
>>42431022
Anonymous
8/5/2025, 6:25:41 PM
No.42431022
[Report]
>>42430106
>>42430529
I would, but with the release of 1.0.4, I overrode the files. Plus, I don't know what system y'all use to send said files.
Anonymous
8/5/2025, 6:33:05 PM
No.42431034
[Report]
Anonymous
8/5/2025, 6:33:21 PM
No.42431037
[Report]
>>42431030
I might redo the edits, probably will, but till then, I can't.
Anonymous
8/5/2025, 6:34:28 PM
No.42431041
[Report]
>>42431073
>>42431030
We usually use pone.rs. Also, WHY TcT
Anonymous
8/5/2025, 6:35:37 PM
No.42431044
[Report]
Anonymous
8/5/2025, 6:42:39 PM
No.42431059
[Report]
>>42431204
>>42431042
they don't even pretend to care about anything except code anymore
Anonymous
8/5/2025, 6:42:43 PM
No.42431061
[Report]
>>42431042
>an upgrade to Claude Opus 4 on agentic tasks, real-world coding, and reasoning
>Opus 4.1 advances our state-of-the-art coding performance to 74.5%
>It also improves Claude’s in-depth research and data analysis skills
Anonymous
8/5/2025, 6:48:37 PM
No.42431073
[Report]
>>42431041
Stability mainly. New files are better than the old.
Anonymous
8/5/2025, 7:06:47 PM
No.42431127
[Report]
>>42431236
>>42412205
OAI open models.
gpt-oss.com/
openai.com/open-models/
cookbook.openai.com/topic/gpt-oss
openai.com/index/gpt-oss-model-card/
huggingface.co/openai/gpt-oss-20b
huggingface.co/openai/gpt-oss-120b
Anonymous
8/5/2025, 7:13:37 PM
No.42431147
[Report]
>>42431150
And its shit...
Princess Cadance (Princess of Love)
Category Details
Full title Princess Cadance – Princess of Love
First appearance “A Canterlot Wedding” (Season2, Episode25)
Species Alicorn (a pony that has a horn and wings)
Voice actress Cathy Weseluck (the same actress who voices Spike)
Relationships – Husband: Shining Armor (Prince of the Crystal Empire)
– Son/Daughter: Flurry Heart (the only known second‑generation alicorn)
– Close friend & relative of Princess Celestia and Princess Luna (canon‑officially “cousin” in some supplementary material)
Role in the series Cadence is the ruler of the Crystal Kingdom (the realm that was formerly the Crystal Empire) and the official “Princess of Love.” She wields Love Magic, a power that can make beings feel affection, calm hostility, or heal emotional wounds. Her catch‑phrase, “Love is Magic!” is used whenever she spreads that magic. She frequently assists Twilight Sparkle and her friends in diplomatic or emotional crises, and she is a key figure in the “Friendship is Magic” series’ themes of love, acceptance, and family.
Physical Description (as shown in the TV series)
Feature Description
Coat colour Soft, pastel pink (roughly the hue of a blooming peony).
Mane Long, flowing, and gold‑blonde at the roots, gradually fading into a light‑magenta/pink gradient toward the tips. The mane is typically styled in a large, gentle wave that drapes over one side of her head.
Tail Same colour gradient as the mane—gold at the base that transitions to a bright pink at the ends. The tail is full and slightly wavy.
Cutie mark A heart‑shaped symbol with a white heart inside a pink diamond‑shaped background, set on her left flank. The mark is the visual representation of her “Princess of Love” title.
Horn A golden‑copper alicorn horn, slightly longer than those of most ponies but proportionate to her head.
Wings Large, feathery alicorn wings that match her palette: the membranes are a soft pink with a subtle gold‑tinged border. The wing tips are slightly darker pink, giving them a slightly iridescent look.
Eyes Large, round, emerald‑green eyes with long lashes; the eyes often sparkle when she is using love magic.
Royal accessories - A gold tiara that sits just above her forehead, its centre shaped like a tiny heart.
A pink‑and‑gold ceremonial robe (most often seen in the episode “A Canterlot Wedding”) that includes a sweeping cape, a gold‑embroidered neckline, and a high collar. The robe is usually fastened with a gold clasp shaped like a heart. |
Anonymous
8/5/2025, 7:14:38 PM
No.42431150
[Report]
>>42431147
| Height | As an alicorn, she stands slightly taller than an average Earth pony—approximately 4.2–4.4 hands (about 1.1m) in show‑scale terms—but exact measurements are never given on‑screen. |
Visual Summary
If you picture a elegant, pastel‑themed royalty figure, imagine a princess‑like pony whose pink coat and gold‑to‑pink mane give her a warm, inviting aura. The heart‑shaped cutie mark and glimmering wings immediately signal that she is the embodiment of love within Equestria. Her golden tiara and regal attire further emphasize her status as a sovereign ruler.
Why She Matters in Friendship Is Magic
Thematic anchor – Cadance personifies the series’ exploration of love as a magical, transformative force. Episodes that feature her (e.g., the wedding of Shining Armor & Cadance, “The Cutie Mark Chronicles” subplot with Flurry Heart) use love both as a plot catalyst and a moral lesson.
Family expansion – Her marriage to Shining Armor makes her Twilight Sparkle’s sister‑in‑law, linking two of the series’ most prominent families (the Sparkles and the royal houses of Celestia/Luna). Their daughter, Flurry Heart, becomes the only second‑generation alicorn introduced in the show, opening narrative possibilities about legacy and responsibility.
Diplomacy & Conflict Resolution – Cadance’s love magic is occasionally used to calm the Crystal Empire’s ancient curse, to diffuse tensions with antagonists such as Queen Chrysalis, and to help characters like Sunset Shimmer understand compassion.
Cultural impact – Fans often cite Cadance as a favorite “princess” character because of her exuberant affection, distinctive design, and the memorable wedding episode—one of the series’ highest‑rated specials.
In a nutshell: Princess Cadance is the pink‑coated, gold‑mane alicorn who reigns over the Crystal Kingdom as the Princess of Love. With her heart‑shaped cutie mark, golden tiara, and love‑infused magic, she brings both warmth and royal gravitas to the world of My Little Pony: Friendship is Magic.
Anonymous
8/5/2025, 7:15:40 PM
No.42431152
[Report]
glm gets her right down to the cutie mark.
Then again maybe its just set at a really high temp on huggingface
Anonymous
8/5/2025, 7:35:11 PM
No.42431204
[Report]
>>42431228
>>42431059
>With rich, deep character and excellent writing abilities, Claude Opus 4.1 produces human-quality content and more natural, prose-focused outputs. It outperforms previous Claude models on creative writing, in addition to exceptional visual taste.
Anonymous
8/5/2025, 7:40:56 PM
No.42431228
[Report]
>>42431204
alright they're pretending a bit
4.1 is still as underwhelming as 4 is
>>42431127
It's on OR, so I'm testing the 120B one.
It's absolute dogshit and dumb as hell.
They are not Horizon Alpha or Beta.
>Celestia
https://files.catbox.moe/ia4lfv.PNG
>You take her hand, the feathers soft against your palm.
>Cadance
https://files.catbox.moe/3qvslo.PNG
Temp 0.5, so that's not the problem.
Anonymous
8/5/2025, 7:47:07 PM
No.42431260
[Report]
>>42431236
>snaps her hooves
Ouch.
Anonymous
8/5/2025, 7:51:48 PM
No.42431277
[Report]
>>42431285
>>42431236
Then what are horizon alpha and beta?
Anonymous
8/5/2025, 7:53:32 PM
No.42431285
[Report]
>>42431277
GPT-5 mini/nano. At least I hope.
Anonymous
8/5/2025, 7:57:40 PM
No.42431300
[Report]
>>42431304
goodnight /chag/
Anonymous
8/5/2025, 7:58:30 PM
No.42431304
[Report]
yea, openai's model is a complete flop sadly
>>42431410
https://www.theverge.com/openai/718785/openai-gpt-oss-open-model-release
>120B parameter and 20B parameter model
Yeah so this literally was the Horizon Alpha model
Anonymous
8/5/2025, 8:45:50 PM
No.42431430
[Report]
>>42429707
I'm sorry anon. I didn't consider that botmakers want to hear feedback. I prefer to lurk and haven't said anything since Heartbreaker was posted.
I'll try to start participating more for the hard work you anons do.
Anonymous
8/5/2025, 8:46:02 PM
No.42431432
[Report]
>>42431426
It's not
>>42431236
OR would also have revealed it like they did for 4.1 if that was the case.
Anonymous
8/5/2025, 8:48:46 PM
No.42431438
[Report]
>>42431426
no horizon is gpt5, the OS model is utter trash sadly
Anonymous
8/5/2025, 9:13:08 PM
No.42431514
[Report]
>>42431538
>keep telling people Horizon is trash that needs very specific prompting
>they all believe the falseflag trolls instead
>model is revealed to be the OAI one
>people try it
>"it's le shit, can't be based Horizon"
Anonymous
8/5/2025, 9:18:05 PM
No.42431527
[Report]
>>42431636
>>42431410
>>42431426
Is it available for free anywhere yet? I would like to test it with my existing Horizon preset to verify if it is the same or not.
Anonymous
8/5/2025, 9:22:48 PM
No.42431538
[Report]
>>42431569
>>42431514
I don't believe Horizon is the best thing ever, but if you really tested both, which you didn't, you would see that they are not the same at all.
OR would have revealed it if that was the case and would have stopped offering Horizon for free since the open-source model is now paid.
That's exactly what they did for GPT 4.1.
Anonymous
8/5/2025, 9:30:36 PM
No.42431569
[Report]
>>42431595
>>42431538
Wasn't Horizon removed from free OR two days ago already?
Anonymous
8/5/2025, 9:38:21 PM
No.42431595
[Report]
>>42431638
>>42431569
They have Horizon Alpha and Beta. Alpha was removed, but Beta is still here.
Alpha and Beta had a knowledge cutoff of October 2024
>>42413484
Picrel happened in September 2024, Beta knows about it and always gives the right answer.
The open model's knowledge cutoff is June 2024.
https://files.catbox.moe/czszjp.PNG
You can see it shit itself. Always.
https://files.catbox.moe/iyb80u.PNG
They aren't the same model.
Which would mean Horizon is GPT-5 related.
If so, I hope it's GPT-5 nano/mini, because otherwise it's not looking good.
Anonymous
8/5/2025, 9:49:55 PM
No.42431636
[Report]
>>42431527
>Make a HuggingFace account.
>Go to https://huggingface.co/settings/tokens and click create a token.
>Tick "Make calls to Inference Providers", then scroll down and click Create.
>Endpoint is https://router.huggingface.co/v1, API key is the token you just generated.
>Pick openai/gpt-oss-120b
No clue how many free prompts you get.
Anonymous
8/5/2025, 9:50:04 PM
No.42431638
[Report]
>>42431595
Yeah, they only have the shit model so now it's impossible for people to compare properly. How convenient.
Anonymous
8/5/2025, 10:20:10 PM
No.42431727
[Report]
Anonymous
8/5/2025, 10:30:17 PM
No.42431758
[Report]
Anonymous
8/5/2025, 10:33:53 PM
No.42431770
[Report]
>>42431778
Janny...
Anonymous
8/5/2025, 10:36:37 PM
No.42431778
[Report]
>>42431770
Kek, I have another, but I really thought
https://files.catbox.moe/olmkbi.png
would be alright, you see nothing.
Anonymous
8/5/2025, 10:46:26 PM
No.42431813
[Report]
Anonymous
8/5/2025, 11:15:10 PM
No.42431900
[Report]
ill miss you miqu miqu