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 Sonnet 4.6 vs GPT-5

Verdict

Claude Sonnet 4.6 leads 7–5 across 12 shared benchmarks.

Claude Sonnet 4.6 7 · GPT-5 5 · higher isn't always better — see caveats.

Per-benchmark head-to-head

Claude Sonnet 4.6GPT-5
  • 23.7
    APEX-Agents
    18.3
  • 86.5
    ARC-AGI
    65.7
  • 60.42
    ARC-AGI-2
    9.86
  • 13
    Chess Puzzles
    37
  • 3.14
    CritPt
    12.6
  • 54.87
    DeepResearch Bench
    55.13
  • 83.16
    GPQA diamond
    81.57
  • 85.78
    OTIS Mock AIME 2024-2025
    91.38
  • 29
    SimpleQA Verified
    50.6
  • 75.21
    SWE-Bench verified
    73.55
  • 53.4
    Terminal Bench
    49.6
  • 66.07
    WeirdML
    60.7
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 Sonnet 4.686.5
  • GPT-598.13