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 4.8 vs GLM-5.2

Verdict

Claude Opus 4.8 leads 11–2 across 14 shared benchmarks.

Claude Opus 4.8 11 · GLM-5.2 2 · 1 tied · higher isn't always better — see caveats.

Per-benchmark head-to-head

Claude Opus 4.8GLM-5.2
  • 42.5
    APEX-Agents
    35.6
  • 92.5
    ARC-AGI
    77
  • 72.08
    ARC-AGI-2
    22.78
  • 34
    Chess Puzzles
    21
  • 20.86
    CritPt
    20.86
  • 63.8
    CursorBench
    54.6
  • 56.1
    FrontierMath-Tier-4-v2-Private
    29.27
  • 80
    FrontierMath-Tiers-1-3-v2-Private
    59.21
  • 70.9
    GBAEval
    0.02
  • 88.05
    GPQA diamond
    89.14
  • 98.33
    OTIS Mock AIME 2024-2025
    86.38
  • 34.08
    PostTrainBench
    34.29
  • 39.5
    SimpleQA Verified
    38.1
  • 82.89
    WeirdML
    70.12
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 4.898.33
  • GLM-5.289.14

GLM-5.2 vs …

  • vs PowerMoE-3b
  • vs Claude Fable 5
  • vs GPT-5.5
  • vs Gemini 3.1 Pro
  • vs Llama 3.1-405B