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 3.5 Flash

Google DeepMind·released 2026-05-19

independently reproduced

Scored on 15 benchmarks, leading 1. 47% of its scores are independently reproduced — the rest are self-reported or unverified.

15
Benchmarks scored
1
SOTA leads
95.55
Best score
47%
Reproduced
7 of 15

Leads on: APEX-Agents

Benchmark fingerprint

Every score, ranked, colored by evidence status.

  • OTIS Mock AIME 2024-2025
    95.55
  • ARC-AGI
    92.5
  • GPQA diamond
    90.4
  • SWE-Bench verified
    79.34
  • ARC-AGI-2
    72.08
  • SimpleBench
    72.04
  • SimpleQA Verified
    68.4
  • FrontierMath-Tiers-1-3-v2-Private
    62.81
  • WeirdML
    62.64
  • Chess Puzzles
    50
  • CursorBench
    49.8
  • APEX-Agents
    49.6
  • FrontierMath-Tier-4-v2-Private
    26.83
  • CritPt
    13.14
  • GBAEval
    6.7

All benchmark scores

BenchmarkScoreEvidence statusMeasured
OTIS Mock AIME 2024-202595.55accuracy (%)reproduced· optimized2026-05-19
ARC-AGI92.5accuracy (%)unverified· optimized2026-05-19
GPQA diamond90.4accuracy (%)reproduced· optimized2026-05-19
SWE-Bench verified79.34accuracy (%)reproduced· optimized2026-05-19
ARC-AGI-272.08accuracy (%)unverified2026-05-19
SimpleBench72.04accuracy (%)unverified2026-05-19
SimpleQA Verified68.4accuracy (%)reproduced2026-05-19
FrontierMath-Tiers-1-3-v2-Private62.81accuracy (%)reproduced2026-05-19
WeirdML62.64accuracy (%)unverified2026-05-19
Chess Puzzles50accuracy (%)reproduced2026-05-19
CursorBench49.8accuracy (%)unverified2026-05-19
APEX-Agents49.6accuracy (%)unverified2026-05-19
FrontierMath-Tier-4-v2-Private26.83accuracy (%)reproduced2026-05-19
CritPt13.14accuracy (%)unverified2026-05-19
GBAEval6.7accuracy (%)unverified2026-05-19
15 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 Flash92.77
  • Gemini 2.5 Pro (Jun 2025)91.7
  • Gemini 3 Pro91.38

Benchmarks it leads

  • APEX-Agents