TAU2
Measured May 14, 2026Source
Score
0.2
Models15 0 
Devstral Medium
MistralMistralFrance
Devstral Medium is a mid-sized model from Mistral's Devstral family, specifically optimized for code generation and software engineering tasks. It balances strong coding and reasoning capabilities with efficient performance and cost-effectiveness.
CodingReasoningFastCheapLong context
Input / 1M tokens
$0.40Output / 1M tokens
$2.00Output tokens/s
83.48First-token seconds
0.51sSupported plans
0Benchmark history
Evaluations
Terminalbench Hard
Measured May 14, 2026Source
Score
0.09
Lcr
Measured May 14, 2026Source
Score
0.29
Ifbench
Measured May 14, 2026Source
Score
0.3
Aime 25
Measured May 14, 2026Source
Score
0.05
Aime
Measured May 14, 2026Source
Score
0.07
Math 500
Measured May 14, 2026Source
Score
0.71
Scicode
Measured May 14, 2026Source
Score
0.29
Livecodebench
Measured May 14, 2026Source
Score
0.34
Hle
Measured May 14, 2026Source
Score
0.04
Gpqa
Measured May 14, 2026Source
Score
0.49
Mmlu Pro
Measured May 14, 2026Source
Score
0.71
Artificial Analysis Math Index
Measured May 14, 2026Source
Score
4.7
Artificial Analysis Coding Index
Measured May 14, 2026Source
Score
15.9
Artificial Analysis Intelligence Index
Measured May 14, 2026Source
Score
18.7
Plan availability
Products and plans that support this model
No products or plans have been linked to this model yet.
Discussion
Thinking... Make sure you are connected to GitHub server