Powered by Replicate and Mistral AI. Run Mixtral in the cloud.

Mistral 8x7B is a high-quality mixture of experts model with open weights, created by Mistral AI. It outperforms Llama 2 70B on most benchmarks with 6x faster inference, and matches or outputs GPT3.5 on most benchmarks.

Mixtral can , write and , , or even

Send me a message.