Magistral Medium 1
Jun 2025
Magistral Medium 1Compare inon
Benchmarks of Magistral Medium 116
BenchmarkByScoreRank
Overall1
Knowledge4
Reasoning4
Science2
Coding1
Legacy · saturated or no longer reported4
No benchmark matches
—per task?
not available via APIprice
—output speed?
40Kcontext
Rankingin thinking mode?
Alternatives?
ModelScoreCostPer taskWhat changes
BetterGPT-6 Lunalow+17.8—$0.005Function calling only with none · Chat Completions+3
OpenMuse Spark Openxhigh+51.6—$0.84Reasoning off leaves output unchanged · Self-hosted / llama.cpp+1
BetterGranite 4.2 3BthinkingOpen+12.2—$0.006
BetterGemma 4 E4BthinkingOpen+7.2—$0.008BetterGPT-5.6 Lunalow+12.3—$0.01ModelScoreCostPer taskWhat changes
Nothing here yet.
ModelScoreCostPer taskWhat changes
GPT-6 Lunalow+17.8—$0.005Function calling only with none · Chat Completions+3Granite 4.2 3BthinkingOpen+12.2—$0.006Gemma 4 E4BthinkingOpen+7.2—$0.008GPT-5.6 Lunalow+12.3—$0.01
MiMo v2.5thinkingOpen+17.1—$0.019ModelScoreCostPer taskWhat changes
Muse Spark OpenxhighOpen+51.6—$0.84Reasoning off leaves output unchanged · Self-hosted / llama.cpp+1
Kimi K3maxOpen+46.6—$2.00Sampling parameters are fixed+3
Step 5 PreviewthinkingOpen+43.3—$1.03web_search not available · Chat Completions+2
GLM 5.3maxOpen+41.2—$2.01Reasoning cannot be disabled+3MMotif 3Open+37.1—$0.50 / $2ModelScoreCostPer taskWhat changes
Mistral Large 4thinking+21.6—$1.13Infinite whitespace stream+4Mistral Small 4thinkingOpen-3.5—$0.015Devstral Mediumnon-thinkingOpen-4.6—$0.40 / $2Mistral Medium 3.5highOpen-9.9—$0.50Mistral Large 3non-thinkingOpen-12.6—$0.031Magistral Small 1thinkingOpen-17.5—$0.70 / $1.16Ministral 3 3B 2512Open-37.1—$0.008Ministral 3 14B 2512non-thinkingOpen-39.4—$0.019All Mistral models →History
2 versionsBenchmarks
all categoriesChat & Knowledge
Humanity's Last Exam9.8#196
AA Omniscience Index-26.9#141
CritPt (physics)0.3#149
AA Intelligence Index 4.3.2≈9.1#228
AA Omniscience · Accuracy20.3#151
GPQA Diamond67.9#198
AIME 2025 · AA archived40.3#96
Coding
LiveCodeBench · AA archived52.7#82
Agents & Tool Use
τ²-bench · AA23.1#179
Writing & Language
IFBench (instructions)25.1#232
FAQ
How much does Magistral Medium 1 cost?›
There is no per-unit API price: not available via API.
Is Magistral Medium 1 free to use?›
There is no free API.
How good is Magistral Medium 1, and what is it best at?›
#195 of 334 in Chat & Knowledge, #128 of 239 in Coding, #177 of 242 in Agents & Tool Use, #229 of 245 in Writing & Language.
Is Magistral Medium 1 better than Kimi K2 0905?›
On AA Omniscience Index, Magistral Medium 1 scores -26.9 vs -26.6 — about the same.
What is the context window of Magistral Medium 1?›
40K tokens.
When was Magistral Medium 1 released? Is there a newer version?›
Released 2025-06-10. It has been superseded by Magistral Medium 1.2.
Is Magistral Medium 1 open source?›
No, Magistral Medium 1 is a closed model available through an API. The closest open-weights option here is Muse Spark Open.
Can Magistral Medium 1 read images?›
No, it accepts text only.
What is Magistral Medium 1?›
Magistral Medium 1 is a closed language model from Mistral, released Jun 2025.
What are the best alternatives to Magistral Medium 1?›
GPT-6 Luna (better), Muse Spark Open (open weights).