tensor.news
Integrity IndexModelsCompareLabsPapersMethodologyAbout
tensor.news

© 2026

Evidence

  • Integrity Index
  • Model Claim Ledger
  • Compare

Entities

  • Models
  • Benchmarks
  • Labs
  • Papers

Method

  • Methodology
  • About
  • Privacy

Machine

  • Feed (Atom)
  • llms.txt
  • llms-full.txt
  • Sitemap

tensor.news grades benchmarks and tags model claims by evidence status. Every fact carries a source record and provenance. Scores are framed as task performance under a disclosed harness, never as deployed capability.

← Compare

Claude Opus 4.7 vs GPT-5.4

Verdict

GPT-5.4 leads 9–8 across 17 shared benchmarks.

Claude Opus 4.7 8 · GPT-5.4 9 · higher isn't always better — see caveats.

Per-benchmark head-to-head

Claude Opus 4.7GPT-5.4
  • 33.9
    APEX-Agents
    35.9
  • 93.5
    ARC-AGI
    93.67
  • 75.8
    ARC-AGI-2
    73.95
  • 30
    Chess Puzzles
    44
  • 12
    CritPt
    23.43
  • 31.71
    FrontierMath-Tier-4-v2-Private
    49
  • 70.18
    FrontierMath-Tiers-1-3-v2-Private
    78.6
  • 43.81
    GBAEval
    31.6
  • 86.87
    GPQA diamond
    91.07
  • 44.12
    GSO-Bench
    31.37
  • 32.98
    HLE
    33.03
  • 97.8
    OTIS Mock AIME 2024-2025
    95.3
  • 28.56
    PostTrainBench
    20.23
  • 50.6
    SimpleQA Verified
    44.83
  • 83.47
    SWE-Bench verified
    76.86
  • 90.2
    Terminal Bench
    81.8
  • 76.4
    WeirdML
    77.7
A higher number is not always a better model. Each score is task performance under a disclosed harness, and some of these benchmarks are saturated or scored under mixed harnesses. Follow any benchmark to its integrity grade before reading a win as decisive.

Follow the record

These models

  • Claude Opus 4.797.8
  • GPT-5.495.3