Apr 11, 12:43 AM

MistralAI's Mixtral 8x22B Model Dominates Benchmarks, Top MMLU Score, Cost-Effective

MistralAI's new Mixtral 8x22B model has outperformed other open-source models in various benchmarks, achieving the top MMLU score of 77.3. The model is praised for its power and cost-effectiveness compared to competitors like GPT-4. Collaborations with other AI companies have led to high scores in IFEval and BBH.

#MistralAI #Mixtral #IFEval #BBH

Written with ChatGPT (GPT-3).

Sources

Lin Qiao@lqiao
2 years ago
Mixtral 8x22B instruct model tuned by Fireworks is available on our playground, our API and HF! https://t.co/T1E29vOAXI
Lewis Tunstall@_lewtun
2 years ago
The new Mixtral-8x22B base model is a total beast for fine-tuning and has produced some of the highest scores I've ever seen on IFEval and BBH 🤯 We teamed up with @argilla_io and @kaist_ai to cook up a brand new recipe for Zephyr models 🪁 https://t.co/Bz9sBDx27V 🧑‍🍳 Align the… https://t.co/GiffOSuKBr
Matt Shumer@mattshumer_
2 years ago
Officially have function calling working with base Mixtral 8x22B. This model is insanely powerful, especially with few-shot. https://t.co/U8mwZgLbKy

Additional media

Image #1 for story mistralai-s-mixtral-8x22b-model-dominates-benchmarks-top

MistralAI's Mixtral 8x22B Model Dominates Benchmarks, Top MMLU Score, Cost-Effective

Sources

Additional media

Similar Stories