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.

SpecificationGPT-4o (2024-11-20)OpenAIClaude Opus 5Anthropic
ReleasedNov 20, 2024Jul 24, 2026
Context window128K1M
Max output16K128K
Input / 1M$2.5$5
Output / 1M$10$25
Cost per task$0.38$2.25
Arena Elo9471,393
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

  • GPT-4o (2024-11-20)
  • Claude Opus 5

  • Intelligence Index

    GPT-4o (2024-11-20)
    11.1
    Claude Opus 5
    63.1

  • Coding Index

    GPT-4o (2024-11-20)not measured
    Claude Opus 5
    78.0

  • Agentic Index

    GPT-4o (2024-11-20)not measured
    Claude Opus 5
    59.2

  • Output speed

    GPT-4o (2024-11-20)
    70 t/s
    Claude Opus 5
    87 t/s

  • Context window

    GPT-4o (2024-11-20)
    128K
    Claude Opus 5
    1M

  • Latency · lower is better

    GPT-4o (2024-11-20)
    546ms
    Claude Opus 5
    1.12s

  • Blended price / 1M · lower is better

    GPT-4o (2024-11-20)
    $4.38
    Claude Opus 5
    $10

  • Cost per task · lower is better

    GPT-4o (2024-11-20)
    $0.38
    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
MetricGPT-4o (2024-11-20)Claude Opus 5
Intelligence Index11.163.1
Coding Index—78.0
Agentic Index—59.2
Output speed70 t/s87 t/s
Context window128K1M
Latency · lower is better546ms1.12s
Blended price / 1M · lower is better$4.38$10
Cost per task · lower is better$0.38$2.25

Evaluation scores

Percentage correct on a common 0–100% scale

  • GPT-4o (2024-11-20)
  • Claude Opus 5

  • GPQA Diamond

    GPT-4o (2024-11-20)
    54.3%
    Claude Opus 5
    93.2%

  • Humanity's Last Exam

    GPT-4o (2024-11-20)
    2.4%
    Claude Opus 5
    54.9%

  • SciCode

    GPT-4o (2024-11-20)
    33.3%
    Claude Opus 5
    55.7%

  • τ²-bench

    GPT-4o (2024-11-20)
    25.1%
    Claude Opus 5
    42.1%

  • Terminal-Bench Hard

    GPT-4o (2024-11-20)
    8.3%
    Claude Opus 5not measured

  • LiveCodeBench

    GPT-4o (2024-11-20)
    30.9%
    Claude Opus 5not measured

  • AA-LCR long context

    GPT-4o (2024-11-20)
    0.0%
    Claude Opus 5
    75.7%

  • AIME 2025

    GPT-4o (2024-11-20)
    6.0%
    Claude Opus 5not measured

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

View as table
EvaluationGPT-4o (2024-11-20)Claude Opus 5
GPQA Diamond54.3%93.2%
Humanity's Last Exam2.4%54.9%
SciCode33.3%55.7%
τ²-bench25.1%42.1%
Terminal-Bench Hard8.3%—
LiveCodeBench30.9%—
AA-LCR long context0.0%75.7%
AIME 20256.0%—

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