#ModelS
Just because you can upload 10 terabytes of poorly optimized ai generated models doesn't mean you should upload 10 terabytes of poorly optimized ai generated models

As a reminder we have had an AI spam clause in our quality guidelines for over two years: itch.io/docs/creator...
Content creator quality guidelines
itch.io
November 3, 2026 at 3:33 AM
Two models off the clock
November 2, 2026 at 9:09 PM
models currently showing 77% odds of you deserving a little treat
November 3, 2026 at 8:56 PM
November 3, 2026 at 9:45 PM
November 4, 2026 at 9:30 AM
Lawn mowers go through more mandatory safety testing than AI models right now, and that’s just nuts.
October 31, 2026 at 6:07 PM
the OG CTR models honestly blend in pretty well with MKDS tracks lol
November 2, 2026 at 2:36 PM
This means that advanced frontier models can directly alter model weights to improve smaller models without pretraining, reinforcement learning, or other fine tuning methods. This is a game changing step in how models can be developed and changed.
November 3, 2026 at 2:13 AM
Leash
Models by Madruga3D
October 30, 2026 at 4:48 PM
Rubicon Models 1/56 A34 Comet #Histscalemodels
November 4, 2026 at 4:10 PM
1 year ago today with sig models!
November 3, 2026 at 1:59 AM
coding models : inexpensive digital cameras :: Steam : YouTube
perhaps; I think we will see to games what happened to the screen with YouTube et al: huge quantities of low effort uninteresting crap that no one ever sees, behind a renaissance of original content by individuals that never could have happened before
November 2, 2026 at 7:23 PM
Trump is going to be playing with his models even more now
November 4, 2026 at 3:43 PM
its important to have role models
November 4, 2026 at 3:02 AM
Here is a mass attempted replication of old findings of LLMs with new models. Much conventional wisdom is outdated.

Recent models no longer invent citations and both hallucination rates & stereotyping has dropped. But they are still sycophantic & lean left. zou-group.github.io/llm-claims-s...
November 3, 2026 at 2:38 AM
November 2, 2026 at 9:11 AM
here i am training small models using data like a chump

“make it better at math” holy fuck
November 2, 2026 at 7:34 PM
"Franzi's Models - Special Edition"

last of the series
November 2, 2026 at 5:04 PM
November 3, 2026 at 10:48 PM
my interest in the ocarina remake is mostly "looking forward to people recreating Captain Scarlet clips with the extracted models"
your first big assignment
November 2, 2026 at 11:36 PM
Models by MrKipler1 on X
November 3, 2026 at 6:33 AM
As an extra bonus, his wife @isalousberg.bsky.social who also created models but in a more cozy style which do still work great with Kay's models so there's more than enough to mix-and-match.

tinytreats.itch.io
November 1, 2026 at 6:39 PM
A 2024 EMNLP paper, On the Relationship between Truth and Political Bias in Language Models, found that training reward models to favor truthfulness tended to produce left leaning political bias. Its fairly obvious thats because left leaning is just emperically correct.
November 3, 2026 at 7:12 PM
Found some time to build some models.
November 2, 2026 at 11:46 PM
currently at the stopping point for work models and personal models which is "i have to model and texture clothes and i dont want to"
November 3, 2026 at 11:52 PM