it would be a dream come true for any users and usage,because a smart ai is more immersive but right now the modele as dumb safety protocole way too strict to even have fun with the model.
My understanding is that venice is primarily a host for various models that others have created, not a model creator itself. As such, it seems to me that any in-built censorship reflex is attributable to the actual creator of the model in question. I think Venice implements something akin to a jailbreak system prompt to such models in order to reduce that reflex, but there's only so much a system prompt can do. Anyway, I'm just saying that Im pretty sure that the problem you're describing is not the fault of Venice AI.
Amaranth Echo
•
Feb 6
si
Jack
Team•
Aug 15, 2025
We are working on improvements here.
Harlequin Zipper
•
Nov 28, 2025
Vision. Reasoning. Not dumb as a brick. Larger context window/filesize support, or perhaps context management for each chat released with it (file uploads) and better context management in general across API and in chats. This would all be so very nice. I love this service and I bagged up on the crypto NGL but there’s no balance. It’s GLM which is stupider than advertised but not the absolute worst, you’re retiring your only semi-worthy reasoning model (lord I hope it’s for a good release) but even it had some severe flaws, and then there’s itty bitty models that someone in my shoes doesn’t even care to use period, API or Chat.
Keep us posted. It would be nice if we had a “Haiku”, “Sonnet”, “Opus” thing going on here, even better if all my other complaints were nicked AND we got special 1M context window increases, etc.
Log in to comment and vote
Comments4
Sapphire Pinwheel
Feb 17
My understanding is that venice is primarily a host for various models that others have created, not a model creator itself. As such, it seems to me that any in-built censorship reflex is attributable to the actual creator of the model in question. I think Venice implements something akin to a jailbreak system prompt to such models in order to reduce that reflex, but there's only so much a system prompt can do. Anyway, I'm just saying that Im pretty sure that the problem you're describing is not the fault of Venice AI.
Amaranth Echo
Feb 6
si
Jack
Aug 15, 2025
We are working on improvements here.
Harlequin Zipper
Nov 28, 2025
Vision. Reasoning. Not dumb as a brick. Larger context window/filesize support, or perhaps context management for each chat released with it (file uploads) and better context management in general across API and in chats. This would all be so very nice. I love this service and I bagged up on the crypto NGL but there’s no balance. It’s GLM which is stupider than advertised but not the absolute worst, you’re retiring your only semi-worthy reasoning model (lord I hope it’s for a good release) but even it had some severe flaws, and then there’s itty bitty models that someone in my shoes doesn’t even care to use period, API or Chat.
Keep us posted. It would be nice if we had a “Haiku”, “Sonnet”, “Opus” thing going on here, even better if all my other complaints were nicked AND we got special 1M context window increases, etc.