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.

← Models

GPT-5.5

OpenAI·released 2026-04-23

independently reproduced

Scored on 22 benchmarks, leading 4. 32% of its scores are independently reproduced — the rest are self-reported or unverified.

22
Benchmarks scored
4
SOTA leads
100
Best score
32%
Reproduced
7 of 22

Leads on: ARC-AGI-2, CL-bench Life, ExploitBench, OTIS Mock AIME 2024-2025

Benchmark fingerprint

Every score, ranked, colored by evidence status.

  • OTIS Mock AIME 2024-2025
    100
  • ARC-AGI
    95
  • GPQA diamond
    92
  • FrontierMath-Tiers-1-3-v2-Private
    85.26
  • ARC-AGI-2
    85
  • WeirdML
    84.91
  • Terminal Bench
    84.7
  • SWE-Bench verified
    80.58
  • FrontierMath-Tier-4-v2-Private
    72.5
  • CursorBench
    64.3
  • SimpleQA Verified
    63.1
  • SimpleBench
    62.8
  • DeepResearch Bench
    54.01
  • Chess Puzzles
    54
  • GBAEval
    53.22
  • ExploitBench
    41
  • GSO-Bench
    40.2
  • APEX-Agents
    38.4
  • CritPt
    27.14
  • PostTrainBench
    25.02
  • CL-bench Life
    22.2
  • OSWorld 2.0
    13

All benchmark scores

BenchmarkScoreEvidence statusMeasured
OTIS Mock AIME 2024-2025100accuracy (%)reproduced· optimized2026-04-23
ARC-AGI95accuracy (%)unverified· optimized2026-04-23
GPQA diamond92accuracy (%)reproduced· optimized2026-04-23
FrontierMath-Tiers-1-3-v2-Private85.26accuracy (%)reproduced2026-04-23
ARC-AGI-285accuracy (%)unverified2026-04-23
WeirdML84.91accuracy (%)unverified2026-04-23
Terminal Bench84.7accuracy (%)unverified· optimized2026-04-23
SWE-Bench verified80.58accuracy (%)reproduced· optimized2026-04-23
FrontierMath-Tier-4-v2-Private72.5accuracy (%)reproduced2026-04-23
CursorBench64.3accuracy (%)unverified2026-04-23
SimpleQA Verified63.1accuracy (%)reproduced2026-04-23
SimpleBench62.8accuracy (%)unverified2026-04-23
DeepResearch Bench54.01accuracy (%)unverified2026-04-23
Chess Puzzles54accuracy (%)reproduced2026-04-23
GBAEval53.22accuracy (%)unverified2026-04-23
ExploitBench41accuracy (%)unverified2026-04-23
GSO-Bench40.2accuracy (%)unverified2026-04-23
APEX-Agents38.4accuracy (%)unverified2026-04-23
CritPt27.14accuracy (%)unverified2026-04-23
PostTrainBench25.02accuracy (%)unverified2026-04-23
CL-bench Life22.2accuracy (%)unverified2026-04-23
OSWorld 2.013accuracy (%)unverified2026-04-23
22 source records · epoch_benchmarks. Every score is evidence-linked.

Follow the record

More from OpenAI

  • GPT-5.5 Pro100
  • GPT-598.13
  • GPT-5 mini97.85
  • o4-mini97.83
  • o397.77
  • o3-pro97.2

Head-to-head

  • vs PowerMoE-3b
  • vs Claude Fable 5
  • vs Gemini 3.1 Pro
  • vs Llama 3.1-405B
  • vs phi-3-small 7.4B
  • vs DeepSeek-V3

Benchmarks it leads

  • ARC-AGI-2
  • CL-bench Life
  • ExploitBench
  • OTIS Mock AIME 2024-2025