Mistral: Mixtral 8x22B Instruct — pricing & benchmarks

By mistralai · Updated 2026-10-06 · Open weights · apache-2.0open weights / self-hostable

What Mistral: Mixtral 8x22B Instruct is used for

Mistral: Mixtral 8x22B Instruct is best suited for open weights / self-hostable. Access it through one Appmo API key alongside 458 other models — Appmo's optimizer can route to it automatically when it is the best value for a request. Weights: huggingface.co/mistralai/Mixtral-8x22B-Instruct-v0.1.

Use it via the Appmo API (OpenAI-compatible)

curl https://appmo.com/v1/chat/completions \
  -H "Authorization: Bearer $APPMO_KEY" \
  -d '{"model":"mistralai/mixtral-8x22b-instruct","messages":[{"role":"user","content":"Hello"}]}'

FAQ

How much does Mistral: Mixtral 8x22B Instruct cost?
Mistral: Mixtral 8x22B Instruct costs $2 per 1M input tokens and $6 per 1M output tokens ($3 blended), as of 2026-10-06. Prices are expected to fall as competition increases.

What is Mistral: Mixtral 8x22B Instruct best for?
Mistral: Mixtral 8x22B Instruct is best suited for open weights / self-hostable.

How do I use Mistral: Mixtral 8x22B Instruct?
Send requests to Appmo's OpenAI-compatible endpoint with model id "mistralai/mixtral-8x22b-instruct" and one API key — no separate provider account needed.

Is Mistral: Mixtral 8x22B Instruct open source?
Yes — its weights are public on Hugging Face (mistralai/Mixtral-8x22B-Instruct-v0.1) under the apache-2.0 license; check the license terms before commercial use.

Can Mistral: Mixtral 8x22B Instruct create images?
No, Mistral: Mixtral 8x22B Instruct works with text.