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.

SpecificationKAT-Coder-Pro V2.5KwaiPilotClaude Opus 5Anthropic
ReleasedJul 10, 2026Jul 24, 2026
Context window256K1M
Max output80K128K
Input / 1M$0.74$5
Output / 1M$2.96$25
Cost per task$0.11$2.25
Arena Elo—1,393
Serving providers15
ParametersUndisclosedUndisclosed
LicenceProprietaryProprietary
Capabilities
  • Tool use
  • 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

  • KAT-Coder-Pro V2.5
  • Claude Opus 5

  • Intelligence Index

    KAT-Coder-Pro V2.5not measured
    Claude Opus 5
    63.1

  • Coding Index

    KAT-Coder-Pro V2.5not measured
    Claude Opus 5
    78.0

  • Agentic Index

    KAT-Coder-Pro V2.5not measured
    Claude Opus 5
    59.2

  • Output speed

    KAT-Coder-Pro V2.5
    54 t/s
    Claude Opus 5
    87 t/s

  • Context window

    KAT-Coder-Pro V2.5
    256K
    Claude Opus 5
    1M

  • Latency · lower is better

    KAT-Coder-Pro V2.5
    1.62s
    Claude Opus 5
    1.12s

  • Blended price / 1M · lower is better

    KAT-Coder-Pro V2.5
    $1.29
    Claude Opus 5
    $10

  • Cost per task · lower is better

    KAT-Coder-Pro V2.5
    $0.11
    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
MetricKAT-Coder-Pro V2.5Claude Opus 5
Intelligence Index—63.1
Coding Index—78.0
Agentic Index—59.2
Output speed54 t/s87 t/s
Context window256K1M
Latency · lower is better1.62s1.12s
Blended price / 1M · lower is better$1.29$10
Cost per task · lower is better$0.11$2.25

Evaluation scores

Percentage correct on a common 0–100% scale

  • KAT-Coder-Pro V2.5
  • Claude Opus 5

  • GPQA Diamond

    KAT-Coder-Pro V2.5not measured
    Claude Opus 5
    93.2%

  • Humanity's Last Exam

    KAT-Coder-Pro V2.5not measured
    Claude Opus 5
    54.9%

  • SciCode

    KAT-Coder-Pro V2.5not measured
    Claude Opus 5
    55.7%

  • τ²-bench

    KAT-Coder-Pro V2.5not measured
    Claude Opus 5
    42.1%

  • AA-LCR long context

    KAT-Coder-Pro V2.5not measured
    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
EvaluationKAT-Coder-Pro V2.5Claude Opus 5
GPQA Diamond—93.2%
Humanity's Last Exam—54.9%
SciCode—55.7%
τ²-bench—42.1%
Terminal-Bench Hard——
LiveCodeBench——
AA-LCR long context—75.7%
AIME 2025——

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