Skip to content
aiqbench

Mistral · model passport

Devstral 2

devstral-2
  • Open weights
  • Modified MIT licence
  • Released 2026-10-13(first seen on LMArena)
4/27axes covered
Best percentile
54.7MMLU-Pro (Artificial Analysis)
Cohort standing
42.8median 60.8

What the data says

Benchmarked on 4 of 27 tracked axes. Every figure below is the model's own published result on that board.

Weakest axes

  • 5.8 pct — WebDev Arena (1,197)

Uneven: 48.9 percentile points between its best and worst axis — 54.7 pct on MMLU-Pro (Artificial Analysis) versus 5.8 pct on WebDev Arena.

  • MMLU-Pro (Artificial Analysis)54.7 pct76.2%Verify →
  • LiveCodeBench (Artificial Analysis)53.1 pct44.8%Verify →
  • IFBench (Artificial Analysis)35.8 pct38.1%Verify →
  • WebDev Arena5.8 pct1,197Verify →
  • OverallNot benchmarked
  • Vision ArenaNot benchmarked
  • Factuality ArenaNot benchmarked
  • Search ArenaNot benchmarked
  • Document ArenaNot benchmarked
  • Agent ArenaNot benchmarked
  • GPQA DiamondNot benchmarked
  • AIME (math)Not benchmarked
  • FrontierMathNot benchmarked
  • SimpleQA VerifiedNot benchmarked
  • ARC-AGI-2Not benchmarked
  • Humanity's Last ExamNot benchmarked
  • SimpleBenchNot benchmarked
  • LiveBenchNot benchmarked
  • BFCL v4 (function calling)Not benchmarked
  • SWE-bench Verified (live board)Not benchmarked
  • Terminal-Bench 4.0Not benchmarked
  • HHEM Hallucination (Vectara)Not benchmarked
  • SWE-bench VerifiedNot benchmarked
  • Terminal-BenchNot benchmarked
  • CodingNot benchmarked
  • Instruction followingNot benchmarked
  • Hard promptsNot benchmarked

Percentiles are the model's standing within each board's own population. Axes marked "not benchmarked" have no published result for this model — that is an absence of data, not a low score.

  • Devstral 2 stands at 42.8 (raw mean 37.3) — measured on 4 of 27 axes.

    standingtemplateVerify →

  • Devstral 2 stands at 42.8 versus a cohort median of 60.8 across 168 models released within 180 days — below the peer median.

    cohorttemplateVerify →

Every claim is a computed fact from the current board data, phrased by a low-temperature model call and rejected by a validation gate if any number is not in the source row. How the leaderboards work →