Magistral Medium
Mistral's enterprise reasoning model - transparent chain-of-thought, closed weights, now superseded.
Retired
Magistral Medium appears to be discontinued or shut down — deprecated May 2026; consolidated into Mistral Medium 3.5 This page is kept as an archive record and is excluded from search engines.
About
Magistral Medium was Mistral AI's enterprise-focused reasoning model, released in June 2025 as the closed counterpart to the open Magistral Small (24B, Apache 2.0). It offered transparent chain-of-thought reasoning across multiple languages, designed for research, strategic planning, and data-driven decision-making, available via Le Chat and the Mistral API.
It received a September 2025 update (Magistral Medium 1.2) before being deprecated on May 22, 2026 - Mistral consolidated its lineup into Mistral Medium 3.5, a dense 128B open-weight model with configurable reasoning effort that retired both Magistral and Devstral 2 as standalone products.
Honest limits: this is a historical entry - the model is discontinued and new integrations are directed to Medium 3.5. Weights were never open. Its interest now is in understanding Mistral's shift from specialized reasoning models to unified hybrid ones.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 6/10
Scored 6/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 88; Trending score 58/100 from researcher assessment.