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-4 TurboOpenAIClaude Opus 5Anthropic
ReleasedApr 9, 2024Jul 24, 2026
Context window128K1M
Max output4K128K
Input / 1M$10$5
Output / 1M$30$25
Cost per task$1.25$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

  • GPT-4 Turbo
  • Claude Opus 5

  • Intelligence Index

    GPT-4 Turbo
    7.7
    Claude Opus 5
    63.1

  • Coding Index

    GPT-4 Turbo
    21.5
    Claude Opus 5
    78.0

  • Agentic Index

    GPT-4 Turbonot measured
    Claude Opus 5
    59.2

  • Output speed

    GPT-4 Turbo
    25 t/s
    Claude Opus 5
    87 t/s

  • Context window

    GPT-4 Turbo
    128K
    Claude Opus 5
    1M

  • Latency · lower is better

    GPT-4 Turbo
    1.03s
    Claude Opus 5
    1.12s

  • Blended price / 1M · lower is better

    GPT-4 Turbo
    $15
    Claude Opus 5
    $10

  • Cost per task · lower is better

    GPT-4 Turbo
    $1.25
    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-4 TurboClaude Opus 5
Intelligence Index7.763.1
Coding Index21.578.0
Agentic Index—59.2
Output speed25 t/s87 t/s
Context window128K1M
Latency · lower is better1.03s1.12s
Blended price / 1M · lower is better$15$10
Cost per task · lower is better$1.25$2.25

Evaluation scores

Percentage correct on a common 0–100% scale

  • GPT-4 Turbo
  • Claude Opus 5

  • GPQA Diamond

    GPT-4 Turbonot measured
    Claude Opus 5
    93.2%

  • Humanity's Last Exam

    GPT-4 Turbo
    3.1%
    Claude Opus 5
    54.9%

  • SciCode

    GPT-4 Turbo
    31.9%
    Claude Opus 5
    55.7%

  • τ²-bench

    GPT-4 Turbonot measured
    Claude Opus 5
    42.1%

  • LiveCodeBench

    GPT-4 Turbo
    29.1%
    Claude Opus 5not measured

  • AA-LCR long context

    GPT-4 Turbonot measured
    Claude Opus 5
    75.7%

  • AIME 2025

    GPT-4 Turbo
    15.0%
    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-4 TurboClaude Opus 5
GPQA Diamond—93.2%
Humanity's Last Exam3.1%54.9%
SciCode31.9%55.7%
τ²-bench—42.1%
Terminal-Bench Hard——
LiveCodeBench29.1%—
AA-LCR long context—75.7%
AIME 202515.0%—

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