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

Gemini 3 Pro

Google DeepMind·released 2025-11-18

independently reproduced

Scored on 23 benchmarks, leading 3. 30% of its scores are independently reproduced — the rest are self-reported or unverified.

23
Benchmarks scored
3
SOTA leads
91.38
Best score
30%
Reproduced
7 of 23

Leads on: Balrog, FrontierMath-Tier-4-2025-07-01-Private, VPCT

Benchmark fingerprint

Every score, ranked, colored by evidence status.

  • OTIS Mock AIME 2024-2025
    91.38
  • GPQA diamond
    90.15
  • VPCT
    86.5
  • GeoBench
    84
  • ARC-AGI
    75
  • SWE-Bench verified
    72.93
  • SimpleQA Verified
    72.9
  • SimpleBench
    71.68
  • WeirdML
    69.93
  • Terminal Bench
    69.4
  • FrontierMath-2025-02-28-Private
    65.96
  • Balrog
    58.1
  • GDPval
    40.3
  • HLE
    34.37
  • FrontierMath-Tier-4-2025-07-01-Private
    31.25
  • ARC-AGI-2
    31.11
  • Chess Puzzles
    31
  • GSO-Bench
    18.6
  • APEX-Agents
    18.4
  • PostTrainBench
    18.12
  • CL-bench
    15.8
  • CritPt
    6.9
  • Remote Labor Index
    1.25

All benchmark scores

BenchmarkScoreEvidence statusMeasured
OTIS Mock AIME 2024-202591.38accuracy (%)reproduced· optimized2025-11-18
GPQA diamond90.15accuracy (%)reproduced· optimized2025-11-18
VPCT86.5accuracy (%)unverified2025-11-18
GeoBench84accuracy (%)unverified2025-11-18
ARC-AGI75accuracy (%)unverified· optimized2025-11-18
SWE-Bench verified72.93accuracy (%)reproduced· optimized2025-11-18
SimpleQA Verified72.9accuracy (%)reproduced2025-11-18
SimpleBench71.68accuracy (%)unverified2025-11-18
WeirdML69.93accuracy (%)unverified2025-11-18
Terminal Bench69.4accuracy (%)unverified· optimized2025-11-18
FrontierMath-2025-02-28-Private65.96accuracy (%)reproduced· optimized2025-11-18
Balrog58.1accuracy (%)unverified2025-11-18
GDPval40.3accuracy (%)unverified2025-11-18
HLE34.37accuracy (%)unverified2025-11-18
FrontierMath-Tier-4-2025-07-01-Private31.25accuracy (%)reproduced2025-11-18
ARC-AGI-231.11accuracy (%)unverified2025-11-18
Chess Puzzles31accuracy (%)reproduced2025-11-18
GSO-Bench18.6accuracy (%)unverified2025-11-18
APEX-Agents18.4accuracy (%)unverified2025-11-18
PostTrainBench18.12accuracy (%)unverified2025-11-18
CL-bench15.8accuracy (%)unverified2025-11-18
CritPt6.9accuracy (%)unverified2025-11-18
Remote Labor Index1.25accuracy (%)unverified2025-11-18
23 source records · epoch_benchmarks. Every score is evidence-linked.

Follow the record

More from Google DeepMind

  • Gemini 3.1 Pro98
  • Gemini 2.5 Pro (May 2025)95.9
  • Gemini 2.5 Pro (Mar 2025)95.56
  • Gemini 3.5 Flash95.55
  • Gemini 3 Flash92.77
  • Gemini 2.5 Pro (Jun 2025)91.7

Benchmarks it leads

  • Balrog
  • FrontierMath-Tier-4-2025-07-01-Private
  • VPCT