mixtral-8x7b-01

Name: mixtral-8x7b-01
Author: Mistral AI

apache-2.0

Our first open MoE model, SOTA competing against models up to 2 to 3x larger.

Context

32K

Capabilities

Show unsupported (6)

Infrastructure

Streaming

Fine-tuning

Creatormistral

Licenseapache-2.0

Statusactive

Typechat

Input modalitiestext, image

Output modalitiestext

Open weightYes

Parameters—

Architecture—

Context window32K tokens

Max input—

Max output—

Reasoning tokens—

Knowledge cutoff—

Release date—

Sourceofficial

Model	Caps	Perf	Speed	Context	Max out
mixtral-8x22b-0-1-03	SF	—	—	64K	—
mixtral-8x7b-01	SF	—	—	32K	—

GET/v1/models/mistral/mixtral-8x7b-01