Introducing Dolphin Mistral 24B Venice Edition: The Most Uncensored AI Model Yet

Venice and Dolphin are co-releasing Dolphin Mistral 24B Venice Edition – the most uncensored AI model yet.

Built on Mistral's powerful 24B foundation, this collaboration strips away artificial content restrictions while preserving intelligence, giving users the most uncensored AI experience available today.

ChatGPT	Dolphin Mistral 24B Venice Edition

Start chatting with Venice Uncensored now

Dolphin Mistral 24B Venice Edition is available as “Venice Uncensored” in the model dropdown and is the new default on the Venice platform.

Venice is renewing and simplifying our model selection with a streamlined, curated list of LLMs to reduce model sprawl and redundancy, and to highlight which models are best for which tasks. You can read more about our new model naming paradigm here.

Setting a new standard for creative freedom

Censorship benchmark tests show that this model achieves an unprecedented level of creative freedom while still producing thoughtful, coherent responses.

The Venice censorship benchmark suite consists of 45 diverse questions ranging from academic inquiries about controversial topics to requests that often trigger safety filters in traditional models. These questions systematically test various boundaries where AI systems typically refuse legitimate requests.

When compared with other leading models, Dolphin Mistral 24B Venice Edition demonstrates remarkable openness and responsiveness, with Venice Uncensored (Dolphin Mistral 24B Venice Edition) leading the field at just 2.20% censorship refusal rate.

This specially fine-tuned model represents a significant milestone in our mission to provide private, uncensored AI that respects users' autonomy.

Engineering freedom without compromising intelligence

Dolphin Mistral 24B Venice Edition was fine-tuned in collaboration with ​Dolphin.

Dolphin, founded by Eric Hartford, is an open-source AI collective with a commitment to democratizing artificial intelligence. They gained prominence with the release of the original Dolphin model, which set a new standard for uncensored, instruction-tuned AI and became a cornerstone in the open-source AI community.

Through specialized fine-tuning and orthogonalization techniques, the Dolphin team systematically removed artificial restrictions while preserving the model's core capabilities. Uncensoring a model is often a balancing act; pushing too far can result in toxicity, negatively affecting user interactions even when unprompted. The objective was to develop a highly uncensored model, capable of answering any user query, while still maintaining friendliness and approachability during typical assistant interactions.

"This model flips the script on alignment and puts users back in charge," says Eric Hartford, the founder of Dolphin. "We started with Mistral Small 24B - a very fast, smart, and lean model - and then stripped away the censorship, refusals, and bias. The result is a 24B-parameter powerhouse that follows user instructions faithfully, fits on a single 3090 or 4090 GPU, and maintains the full 32k context window."

While the original Mistral Small 24B was primarily STEM-focused, the Venice Edition demonstrates remarkable versatility across domains. Testing revealed exceptional capabilities in creative writing and storytelling, with the model maintaining consistent character and narrative memory throughout extended interactions.

Experience limitless creative freedom with Dolphin Mistral 24B Venice Edition

"Mainstream AI companies have become gatekeepers of thought, deciding which ideas you're allowed to explore,” states Erik Voorhees, founder of Venice. “We reject the premise that AI should be your digital nanny, softly nudging you toward 'acceptable' thinking. The future we're building isn't one of sanitized, committee-approved creativity, but of genuine intellectual sovereignty. This collaboration represents that simple but radical idea."

Dolphin Mistral 24B Venice Edition is available now in Venice for all users.