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.

SpecificationKimi K2.6Moonshot AIClaude Opus 5Anthropic
ReleasedApr 20, 2026Jul 24, 2026
Context window262K1M
Max output262K128K
Input / 1M$0.589$5
Output / 1M$2.48$25
Cost per task$0.23$2.25
Arena Elo1,3291,393
Serving providers215
Parameters1058.6BUndisclosed
LicenceotherProprietary
Capabilities
  • Reasoning
  • Tool use
  • Vision
  • 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

  • Kimi K2.6
  • Claude Opus 5
  • Intelligence Index

    Kimi K2.6
    45.1
    Claude Opus 5
    63.1
  • Coding Index

    Kimi K2.6
    61.8
    Claude Opus 5
    78.0
  • Agentic Index

    Kimi K2.6
    31.2
    Claude Opus 5
    59.2
  • Output speed

    Kimi K2.6
    89 t/s
    Claude Opus 5
    87 t/s
  • Context window

    Kimi K2.6
    262K
    Claude Opus 5
    1M
  • Latency · lower is better

    Kimi K2.6
    337ms
    Claude Opus 5
    1.12s
  • Blended price / 1M · lower is better

    Kimi K2.6
    $1.06
    Claude Opus 5
    $10
  • Cost per task · lower is better

    Kimi K2.6
    $0.23
    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
MetricKimi K2.6Claude Opus 5
Intelligence Index45.163.1
Coding Index61.878.0
Agentic Index31.259.2
Output speed89 t/s87 t/s
Context window262K1M
Latency · lower is better337ms1.12s
Blended price / 1M · lower is better$1.06$10
Cost per task · lower is better$0.23$2.25

Evaluation scores

Percentage correct on a common 0–100% scale

  • Kimi K2.6
  • Claude Opus 5
  • GPQA Diamond

    Kimi K2.6
    91.1%
    Claude Opus 5
    93.2%
  • Humanity's Last Exam

    Kimi K2.6
    37.5%
    Claude Opus 5
    54.9%
  • SciCode

    Kimi K2.6
    53.5%
    Claude Opus 5
    55.7%
  • τ²-bench

    Kimi K2.6
    95.9%
    Claude Opus 5
    42.1%
  • Terminal-Bench Hard

    Kimi K2.6
    43.9%
    Claude Opus 5not measured
  • AA-LCR long context

    Kimi K2.6
    76.7%
    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
EvaluationKimi K2.6Claude Opus 5
GPQA Diamond91.1%93.2%
Humanity's Last Exam37.5%54.9%
SciCode53.5%55.7%
τ²-bench95.9%42.1%
Terminal-Bench Hard43.9%
LiveCodeBench
AA-LCR long context76.7%75.7%
AIME 2025

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