Dolphin Mistral 24B Venice Edition
Dolphin Mistral 24B Venice Edition is an open-weight language model from the Dolphin project (dphn, formerly Cognitive Computations), fine-tuned from Mistral Small 24B Instruct…
About Dolphin Mistral 24B Venice Edition
Dolphin Mistral 24B Venice Edition is an open-weight language model from the Dolphin project (dphn, formerly Cognitive Computations), fine-tuned from Mistral Small 24B Instruct 2501. It was built in collaboration with Venice.ai to reduce refusals and give the system owner control over tone and behaviour through the system prompt. The model is distributed on Hugging Face and is used as the default chat model, named Venice Uncensored, on the Venice.ai service.
The model has about 24 billion parameters, is licensed under Apache-2.0 and keeps Mistral's chat template. The authors recommend a low sampling temperature of around 0.15. Community GGUF quantizations reduce the file size to roughly 11-25 GB, so it can run locally in tools such as Ollama, LM Studio or KoboldCpp on a GPU with 16-24 GB of memory. The project also publishes smaller Dolphin models, including a Llama 3.1 8B variant.
User reviews
Loading reviews…
Similar tools
Bring you spiciest sexual fantasies to life with Promptchan AI Porn generator. Enjoy generating nude AI porn in a range of different styles, from Hentai, Realistic, Fury and…
Welcome to PornJourney! The best AI porn images, AI hentai generator and AI porn video. PornJourney's AI is using the most advanced hand-engineered AI porn. Pricing PornJourney:…
FreedomGPT is a web and mobile AI chat service that gives access to a large catalogue of language models from different providers under one account. An Auto mode selects a model…
Hermes 4 is a family of open-weight language models from Nous Research, released in August 2025 as the successor to Hermes 3. The models are post-trained on a large synthetic…