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.

SpecificationM2.7MiniMaxClaude Opus 5Anthropic
ReleasedMar 18, 2026Jul 24, 2026
Context window205K1M
Max output131K128K
Input / 1M$0.27$5
Output / 1M$1.08$25
Cost per task$0.10$2.25
Arena Elo—1,393
Serving providers95
Parameters228.7BUndisclosed
LicenceotherProprietary
Capabilities
  • Reasoning
  • Tool use
  • Open weights
  • 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

  • M2.7
  • Claude Opus 5

  • Intelligence Index

    M2.7
    38.9
    Claude Opus 5
    63.1

  • Coding Index

    M2.7
    52.6
    Claude Opus 5
    78.0

  • Agentic Index

    M2.7
    25.9
    Claude Opus 5
    59.2

  • Output speed

    M2.7
    307 t/s
    Claude Opus 5
    87 t/s

  • Context window

    M2.7
    205K
    Claude Opus 5
    1M

  • Latency · lower is better

    M2.7
    304ms
    Claude Opus 5
    1.12s

  • Blended price / 1M · lower is better

    M2.7
    $0.473
    Claude Opus 5
    $10

  • Cost per task · lower is better

    M2.7
    $0.10
    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
MetricM2.7Claude Opus 5
Intelligence Index38.963.1
Coding Index52.678.0
Agentic Index25.959.2
Output speed307 t/s87 t/s
Context window205K1M
Latency · lower is better304ms1.12s
Blended price / 1M · lower is better$0.473$10
Cost per task · lower is better$0.10$2.25

Evaluation scores

Percentage correct on a common 0–100% scale

  • M2.7
  • Claude Opus 5

  • GPQA Diamond

    M2.7
    87.4%
    Claude Opus 5
    93.2%

  • Humanity's Last Exam

    M2.7
    29.6%
    Claude Opus 5
    54.9%

  • SciCode

    M2.7
    47.0%
    Claude Opus 5
    55.7%

  • τ²-bench

    M2.7
    84.8%
    Claude Opus 5
    42.1%

  • Terminal-Bench Hard

    M2.7
    39.4%
    Claude Opus 5not measured

  • AA-LCR long context

    M2.7
    75.3%
    Claude Opus 5
    75.7%

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

View as table
EvaluationM2.7Claude Opus 5
GPQA Diamond87.4%93.2%
Humanity's Last Exam29.6%54.9%
SciCode47.0%55.7%
τ²-bench84.8%42.1%
Terminal-Bench Hard39.4%—
LiveCodeBench——
AA-LCR long context75.3%75.7%
AIME 2025——

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