We just released Mixtral-8x22B-v0.1 and Mixtral-8x22B-Instruct-v0.1: - Free to use under Apache 2.0 license - Outperforms all open models - Native function calling - Masters English, French, Italian, German and Spanish. - Seq_len = 64K mistral.ai/news/mixtral-8…
Mixtral-8x22B outperforms all open models across reasoning, knowledge, code and math capabilities while offering the best performance/cost ratio. Mixtral-8x22B-v0.1 on HF: huggingface.co/mistralai/Mixt… Mixtral-8x22B-Instruct-v0.1 on HF: huggingface.co/mistralai/Mixt…
@dchaplot It does well my NYT Connections benchmark x.com/lechmazur/stat…
@dchaplot It does well my NYT Connections benchmark x.com/lechmazur/stat…
@dchaplot Congratulations. Didn't expect to get the instruct version publicly so soon. can run 3.75 bpw on 72GB VRAM
@dchaplot Very cool! Let me know if someone wants to come on-air to talk about it. I host two nationally syndicated radio shows about Data and the Information Economy, DM Radio and InsideAnalysis. We'd love to get you some free publicity for this amazing innovation! #LLM #AI #Innovation
@dchaplot Why there's no comparison to Command R+ on multilingual performance?
@dchaplot Excited with your updates and advances 👏 👨💻🧪
@dchaplot 🫡Got It. Many applications are on the way.