AI Radar

Magistral Small 1.1

Available · Language, Reasoning

Magistral Small 1.1 is an update of Magistral Small 1.0, Mistral AI's 24B-parameter reasoning model, released on 24 July 2025 as open weights under Apache 2.0 and as API model magistral-small-2507. The model card describes changes to tone, formatting and reasoning markup, and says performance is about the same as version 1.0. [1] [2]July 24, 2025 Primary source

Timeline of Magistral Small 1.1 →

Claims and evidence

  • Released Primary source[2]July 24, 2025 [3]release date
  • Deprecated Primary source[4]Deprecated table, Magistral Small 1.1 (magistral-small-2507)
  • Retired Other sources give: [5]Model Lifecycle table, Retirement Date The Mistral Legal Center lifecycle table gives October 31, 2025 as retirement date, the date the docs models overview lists as deprecation date. Primary source[4]Deprecated table, Magistral Small 1.1 (magistral-small-2507) [3]Retirement date
  • Status AvailableApache 2.0 weights are still downloadable from Mistral AI's Hugging Face organisation; the API model magistral-small-2507 was retired on 2025-11-30. The Legal Center lifecycle table lists the model as Retired. Primary source[1]
  • Replaced by Mistral Small 4Replacement named for the API model when checked on 2026-10-01. Primary source[4]Deprecated table, Magistral Small 1.1, Alternative [3]Replacement
  • Successor of Magistral Small Primary source[1]Updates compared with Magistral Small 1.0
  • Derived from (fine tune) Mistral Small 3.1Supervised fine-tuning on Magistral Medium traces followed by reinforcement learning, on top of Mistral Small 3.1 (2503). Primary source[1]first paragraph and base_model metadata
  • Change · Other Better tone and behaviour, with improved LaTeX and Markdown formatting and shorter answers to easy prompts.Compared with Magistral Small Primary source[1]
  • Change · Other Less likely to enter infinite generation loops.Compared with Magistral Small Primary source[1]
  • Change · Reasoning Special [THINK] and [/THINK] tokens wrap the reasoning trace, and the reasoning prompt moved into the system prompt.Compared with Magistral Small Primary source[1]
  • Input text Primary source[3]Modalities
  • Output text Primary source[3]Modalities
  • Open weights Yes Primary source[1] [3]Weights table
  • Context window 40k tokensThe Hugging Face card states a 128k window but says performance may degrade past 40k and recommends a 40k maximum. Primary source[3]Context
  • Parameters 24B Primary source[1]first paragraph [3]Weights table, Parameters (B)
  • Access API, Open-weights download Primary source[2]July 24, 2025 [1]

Lineage

Predecessors

Successors

Based on

Variants and derived

None recorded.

Siblings

None recorded.

All ancestors

All descendants

Variants

No variants recorded in this record.

Related AI Radar coverage

AI Radar coverage starts in June 2026; no coverage linked yet.

All model releases from Mistral AI on AI Radar →