Skip to content
llmwaves

Compare

Four models, one page

Pick the models you are actually choosing between. Every row is scaled within itself, so a bar means something next to its neighbours and nothing across rows.

SpecificationMistral Medium 3.1Mistral AIClaude Opus 5Anthropic
ReleasedAug 13, 2025Jul 24, 2026
Context window131K1M
Max output—128K
Input / 1M$0.4$5
Output / 1M$2$25
Cost per task$0.07$2.25
Arena Elo—1,393
Serving providers15
ParametersUndisclosedUndisclosed
LicenceProprietaryProprietary
Capabilities
  • Tool use
  • Vision
  • Reasoning
  • Tool use
  • Vision

Metrics side by side

Each row is scaled to the largest value in that row — bars compare within a row, never across rows

  • Mistral Medium 3.1
  • Claude Opus 5

  • Intelligence Index

    Mistral Medium 3.1
    14.7
    Claude Opus 5
    63.1

  • Coding Index

    Mistral Medium 3.1
    20.5
    Claude Opus 5
    78.0

  • Agentic Index

    Mistral Medium 3.1
    6.1
    Claude Opus 5
    59.2

  • Output speed

    Mistral Medium 3.1
    51 t/s
    Claude Opus 5
    87 t/s

  • Context window

    Mistral Medium 3.1
    131K
    Claude Opus 5
    1M

  • Latency · lower is better

    Mistral Medium 3.1
    844ms
    Claude Opus 5
    1.12s

  • Blended price / 1M · lower is better

    Mistral Medium 3.1
    $0.8
    Claude Opus 5
    $10

  • Cost per task · lower is better

    Mistral Medium 3.1
    $0.07
    Claude Opus 5
    $2.25

Rows marked “lower is better” still draw a longer bar for a larger number — read the value, not just the length. Arena Elo is in the specification table above instead: it has no meaningful zero, so a bar would flatten the gaps.

View as table
MetricMistral Medium 3.1Claude Opus 5
Intelligence Index14.763.1
Coding Index20.578.0
Agentic Index6.159.2
Output speed51 t/s87 t/s
Context window131K1M
Latency · lower is better844ms1.12s
Blended price / 1M · lower is better$0.8$10
Cost per task · lower is better$0.07$2.25

Evaluation scores

Percentage correct on a common 0–100% scale

  • Mistral Medium 3.1
  • Claude Opus 5

  • GPQA Diamond

    Mistral Medium 3.1
    58.8%
    Claude Opus 5
    93.2%

  • Humanity's Last Exam

    Mistral Medium 3.1
    4.7%
    Claude Opus 5
    54.9%

  • SciCode

    Mistral Medium 3.1
    33.8%
    Claude Opus 5
    55.7%

  • τ²-bench

    Mistral Medium 3.1
    40.6%
    Claude Opus 5
    42.1%

  • Terminal-Bench Hard

    Mistral Medium 3.1
    10.6%
    Claude Opus 5not measured

  • LiveCodeBench

    Mistral Medium 3.1
    40.6%
    Claude Opus 5not measured

  • AA-LCR long context

    Mistral Medium 3.1
    21.3%
    Claude Opus 5
    75.7%

  • AIME 2025

    Mistral Medium 3.1
    38.3%
    Claude Opus 5not measured

A missing bar means that evaluation was not run for that model — it is not a zero.

View as table
EvaluationMistral Medium 3.1Claude Opus 5
GPQA Diamond58.8%93.2%
Humanity's Last Exam4.7%54.9%
SciCode33.8%55.7%
τ²-bench40.6%42.1%
Terminal-Bench Hard10.6%—
LiveCodeBench40.6%—
AA-LCR long context21.3%75.7%
AIME 202538.3%—

Estimated from list pricing: 50K input tokens plus 80K output tokens for reasoning models (25K for non-reasoning).