Venice: Uncensored (free)
Venice Uncensored (free) has no benchmark data and a very small 32K context window, making it unsuitable for most professional business applications. Without any performance evidence, it cannot be recommended beyond experimental or personal use.
Assessment date: March 14, 2026
Our methodology takes into account a range of factors including pricing, functionality, capabilities, benchmark performance, and real-world applicability. Rankings are reviewed and updated regularly as new models are released. Issues with our rankings? Contact us
Venice Uncensored Dolphin Mistral 24B Venice Edition is a fine-tuned variant of Mistral-Small-24B-Instruct-2501, developed by dphn.ai in collaboration with Venice.ai. This model is designed as an “uncensored” instruct-tuned LLM, preserving user control over alignment, system prompts, and behavior. Intended for advanced and unrestricted use cases, Venice Uncensored emphasizes steerability and transparent behavior, removing default safety and alignment layers typically found in mainstream assistant models.
Architecture
| Modality | Text → Text |
| Tokenizer | Other |
| Parameters | 24B |
Model Information
Live Performance
Live endpoint metrics — refreshed every 30 minutes.
