Mixtral 8x22B — Mistral AI

Mixtral 8x22B is a mixture-of-experts model from Mistral AI: it activates only some of its experts per request, delivering the quality of a very large model with the efficiency of a smaller one.

Type : Text · Parameters : 8x22B (MoE) · Context : 64K tokens

Strengths

• High-performing MoE architecture
• Multilingual
• Good at code

Best for

• Varied tasks
• Development
• Multilingual writing

Mixtral 8x22B is served via NVIDIA NIM, included in your subscription with dozens of other models — one price, every model.