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

Kimi K2.6

Moonshot AI·released 2026-04-20

independently reproduced

Scored on 14 benchmarks. 50% of its scores are independently reproduced — the rest are self-reported or unverified.

14
Benchmarks scored
0
SOTA leads
96.11
Best score
50%
Reproduced
7 of 14

Benchmark fingerprint

Every score, ranked, colored by evidence status.

  • OTIS Mock AIME 2024-2025
    96.11
  • GPQA diamond
    87.71
  • SWE-Bench verified
    76.65
  • FrontierMath-Tiers-1-3-v2-Private
    57.19
  • WeirdML
    55.86
  • CursorBench
    47.6
  • SimpleQA Verified
    38.7
  • Chess Puzzles
    26
  • FrontierMath-Tier-4-v2-Private
    25.64
  • APEX-Agents
    18.9
  • ExploitBench
    18
  • CritPt
    8
  • OSWorld 2.0
    4.6
  • GBAEval
    0.86

All benchmark scores

BenchmarkScoreEvidence statusMeasured
OTIS Mock AIME 2024-202596.11accuracy (%)reproduced· optimized2026-04-20
GPQA diamond87.71accuracy (%)reproduced· optimized2026-04-20
SWE-Bench verified76.65accuracy (%)reproduced· optimized2026-04-20
FrontierMath-Tiers-1-3-v2-Private57.19accuracy (%)reproduced2026-04-20
WeirdML55.86accuracy (%)unverified2026-04-20
CursorBench47.6accuracy (%)unverified2026-04-20
SimpleQA Verified38.7accuracy (%)reproduced2026-04-20
Chess Puzzles26accuracy (%)reproduced2026-04-20
FrontierMath-Tier-4-v2-Private25.64accuracy (%)reproduced2026-04-20
APEX-Agents18.9accuracy (%)unverified2026-04-20
ExploitBench18accuracy (%)unverified2026-04-20
CritPt8accuracy (%)unverified2026-04-20
OSWorld 2.04.6accuracy (%)unverified2026-04-20
GBAEval0.86accuracy (%)unverified2026-04-20
14 source records · epoch_benchmarks. Every score is evidence-linked.

Follow the record

More from Moonshot AI

  • Kimi K2.7 Code96.39
  • Kimi K2.592.19
  • Kimi K2 (Sep 2025)87.29
  • Kimi K2 (Jul 2025)86.93
  • Kimi K2 Thinking83.04
  • Kimi-K2-Instruct-0905