Mistral Medium 3
May 2025
Retired 2026-08-31Replace withMistral Medium 3.5Newer version
MiMo v2.6 ProCheaper, as good
GPT-6 LunaBetter
Mistral Medium 3.5Newer versionMiMo v2.6 ProCheaper, as goodGPT-6 LunaBetterMistral Medium 3Compare inon
Benchmarks of Mistral Medium 336
BenchmarkByScoreRank
Overall1
Knowledge4
Math2
Science2
Coding1
Multimodal2
Writing15
Documents1
Legacy · saturated or no longer reported8
No benchmark matches
Rankingin non-thinking mode?
Retired models are left out of the current rankings.
Alternatives?
ModelScoreCostPer taskWhat changes
CheaperMiMo v2.6 ProthinkingOpen×5.1-32%$0.13Pass back reasoning_content+3BetterGPT-6 Lunalow×2.4-75%$0.005Function calling only with none · Chat Completions+3
CheaperDeepSeek V4.1 Flash 1017maxOpen×5-34%$0.26Pass back reasoning_content · Chat Completions+4
CheaperClaude Haiku 5.5max×4.8-75%$0.21Remove sampling parameters+5
CheaperMiniMax M3.1 Flashthinking×4.7-67%$0.22Complete model response must be appendedModelScoreCostPer taskWhat changes
MiMo v2.6 ProthinkingOpen×5.1-32%$0.13Pass back reasoning_content+3DeepSeek V4.1 Flash 1017maxOpen×5-34%$0.26Pass back reasoning_content · Chat Completions+4Claude Haiku 5.5max×4.8-75%$0.21Remove sampling parameters+5MiniMax M3.1 Flashthinking×4.7-67%$0.22Complete model response must be appended
GLM 5.3 FlashmaxOpen×4.6-70%$0.28Defaults to max · Self-hosted+4ModelScoreCostPer taskWhat changes
GPT-6 Lunalow×2.4-75%$0.005Function calling only with none · Chat Completions+3GPT-5.6 Lunalow×2.3-44%$0.01Mistral Small 4thinkingOpen+25%-67%$0.015MiMo v2.5thinkingOpen×2.4-78%$0.019
Granite 4.2 8BthinkingOpen+22%-87%$0.024Context: 131K → 128K+1ModelScoreCostPer taskWhat changes
MiMo v2.6 ProthinkingOpen×5.1-32%$0.13Pass back reasoning_content+3DeepSeek V4.1 Flash 1017maxOpen×5-34%$0.26Pass back reasoning_content · Chat Completions+4GLM 5.3maxOpen×4.9+169%$2.01No image input+4
Step 5 PreviewthinkingOpen×4.8+78%$1.03web_search not available · Chat Completions+2
Kimi K3maxOpen×4.8+650%$2.00Sampling parameters are fixed+3ModelScoreCostPer taskWhat changes
Mistral Large 4thinking×4.2—$1.13Infinite whitespace stream+4Mistral Medium 3.5highOpen+57%—$0.50Mistral Small 4thinkingOpen+25%—$0.015Mistral Large 3non-thinkingOpensame—$0.031Magistral Medium 1thinkingsame——Context: 131K → 40K+1Devstral Mediumnon-thinkingOpensame—$0.40 / $2Context: 131K → 128K+1Magistral Small 1thinkingOpensame—$0.70 / $1.16Context: 131K → 128K+1Mistral Small 3non-thinkingOpen-26%—$0.05 / $0.08Context: 131K → 32K+1All Mistral models →History
3 versionsBenchmarks
all categoriesChat & Knowledge
AA Omniscience Index-31.4
GPQA Diamond57.8
Coding
Agents & Tool Use
τ²-bench · AA24.3
Writing & Language
Vision & Documents
FAQ
How much does Mistral Medium 3 cost?›
$0.40 per 1M input tokens and $2 per 1M output tokens; cached input $0.04 per 1M.
Is Mistral Medium 3 free to use?›
There is no free API; the cheapest paid price is $0.40 per 1M input tokens.
How can I use Mistral Medium 3?›
Through the API of Mistral.
What is the context window of Mistral Medium 3?›
131K tokens, with up to 105K output tokens.
When was Mistral Medium 3 released? Is there a newer version?›
Released 2025-05-07. It has been superseded by Mistral Medium 3.5.
Is Mistral Medium 3 open source?›
No, Mistral Medium 3 is a closed model available through an API. The closest open-weights option here is MiMo v2.6 Pro.
Can Mistral Medium 3 read images?›
It accepts text and images, PDF files as input.
What is Mistral Medium 3?›
Mistral Medium 3 is a closed language model from Mistral, released May 2025.
What are the best alternatives to Mistral Medium 3?›
MiMo v2.6 Pro (cheaper), GPT-6 Luna (better).