Mistral Medium 3

Reasoning

—

Knowledge cutoff

May 2025

Family

mistral-medium

Model ID

mistral-medium-2505

Not stated on Mistral AI’s pages: max input, reasoning efforts, reasoning can be off, max thinking budget, built-in tools, aliases.

• Epoch Capabilities Index 132.9

  #111 of 118 listed models · CI 130.1–135.4

• Epoch AI · GPQA Diamond 59.5%

  #112 of 168 listed models · ±2.82

• Epoch AI · MATH Level 5 81.6%

  #21 of 64 listed models · ±0.9

• Epoch AI · OTIS Mock AIME 2024–2025 32.2%

  #118 of 159 listed models · ±5.97

• Epoch AI, 'Capabilities & benchmarking'. Published online at epoch.ai. Retrieved from 'https://epoch.ai/benchmarks'. Licensed CC BY 4.0. · License · as of Oct 18 – Nov 5, 2026
• changes we made

1. Oct 29, 2026 Now priced on Mistral AI API Input $0.40 · Output $2.00

• Models Overview | Mistral Docs (opens in a new tab) docs.mistral.ai · checked Nov 8, 2026
• Changelog | Mistral Docs (opens in a new tab) docs.mistral.ai · checked Nov 9, 2026
• models.dev: mistral/mistral-medium-2505 (community data, MIT) (opens in a new tab) github.com · checked Oct 10, 2026
• Mistral Medium 3 - Mistral AI | Mistral Docs (opens in a new tab) docs.mistral.ai · checked Nov 3, 2026
• Mistral Medium 3 (opens in a new tab) docs.mistral.ai · checked Nov 5, 2026
• models.dev: mistral-medium-2505 (community data, MIT) (opens in a new tab) github.com · checked Oct 29, 2026

Mistral AI’s official policies: data use, privacy, terms →

Some values on this page are still waiting to be checked against official pages: a filled dot is verified, a hollow one is not yet (with or without a source on file). How we verify data

Mistral AI · France · mistral.ai

Snapshots, the same family and other models from this provider

• Mistral Large 4
• Mistral Small 4
• Ministral 3 14B
• Ministral 3 3B
• Ministral 3 8B
• Magistral Medium 1.2
• Mistral Nemo 12B
• Mistral-7B-Instruct-v0.1