Skip to content
← Models
ClosedMistral AI/10 trustVerified

Mistral Medium 3.1

Frontier-class multimodal at $0.40/M input: Mistral's enterprise workhorse.

Visit official page →

About

Mistral Medium 3.1 is Mistral AI's frontier-class multimodal model, released August 12, 2025 (API ID mistral-medium-2508). It processes text and images with a 128K-token context window, supports external tools and APIs, and generates structured data — positioned as the enterprise sweet spot between the cheap Small line and the flagship Large models. At $0.40 per million input tokens and $2.00 output, it undercuts most frontier multimodal APIs while scoring 21.3 on the Artificial Analysis Intelligence Index, well above the median for its price tier.

The honest picture: it's a proprietary API-only model — Mistral's open-weights generosity doesn't extend to the Medium line. It's not a reasoning model, so complex multi-step problems go to Magistral or competitors' thinking models instead. Output speed (82 tok/s) trails the median for its tier, and the 128K context looks modest next to 1M-context rivals. It's a value-frontier play, not a capability-frontier one.

Specs

Context window

Pricing

Free quota

API access

Weights

Closed — API only

Released

August 12, 2025

Benchmarks

Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.

Use cases

multimodalenterprise-aidocument-analysisstructured-data

Trust score — 8/10

/10 trustVerified

Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 47; Trending score 76/100 from researcher assessment.

More in the mistral family

Sources