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.

SpecificationGLM 4.7Z.aiClaude Opus 5Anthropic
ReleasedDec 22, 2025Jul 24, 2026
Context window205K1M
Max output131K128K
Input / 1M$0.4$5
Output / 1M$1.75$25
Cost per task$0.16$2.25
Arena Elo1,2481,393
Serving providers95
Parameters358.3BUndisclosed
LicencemitProprietary
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

  • GLM 4.7
  • Claude Opus 5

  • Intelligence Index

    GLM 4.7
    34.5
    Claude Opus 5
    63.1

  • Coding Index

    GLM 4.7
    45.3
    Claude Opus 5
    78.0

  • Agentic Index

    GLM 4.7
    26.2
    Claude Opus 5
    59.2

  • Output speed

    GLM 4.7
    502 t/s
    Claude Opus 5
    87 t/s

  • Context window

    GLM 4.7
    205K
    Claude Opus 5
    1M

  • Latency · lower is better

    GLM 4.7
    311ms
    Claude Opus 5
    1.12s

  • Blended price / 1M · lower is better

    GLM 4.7
    $0.738
    Claude Opus 5
    $10

  • Cost per task · lower is better

    GLM 4.7
    $0.16
    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
MetricGLM 4.7Claude Opus 5
Intelligence Index34.563.1
Coding Index45.378.0
Agentic Index26.259.2
Output speed502 t/s87 t/s
Context window205K1M
Latency · lower is better311ms1.12s
Blended price / 1M · lower is better$0.738$10
Cost per task · lower is better$0.16$2.25

Evaluation scores

Percentage correct on a common 0–100% scale

  • GLM 4.7
  • Claude Opus 5

  • GPQA Diamond

    GLM 4.7
    85.9%
    Claude Opus 5
    93.2%

  • Humanity's Last Exam

    GLM 4.7
    27.4%
    Claude Opus 5
    54.9%

  • SciCode

    GLM 4.7
    45.1%
    Claude Opus 5
    55.7%

  • τ²-bench

    GLM 4.7
    95.9%
    Claude Opus 5
    42.1%

  • Terminal-Bench Hard

    GLM 4.7
    31.8%
    Claude Opus 5not measured

  • LiveCodeBench

    GLM 4.7
    89.4%
    Claude Opus 5not measured

  • AA-LCR long context

    GLM 4.7
    68.0%
    Claude Opus 5
    75.7%

  • AIME 2025

    GLM 4.7
    95.0%
    Claude Opus 5not measured

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

View as table
EvaluationGLM 4.7Claude Opus 5
GPQA Diamond85.9%93.2%
Humanity's Last Exam27.4%54.9%
SciCode45.1%55.7%
τ²-bench95.9%42.1%
Terminal-Bench Hard31.8%—
LiveCodeBench89.4%—
AA-LCR long context68.0%75.7%
AIME 202595.0%—

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