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

Grok 4.3 Beta

xAI·released 2026-04-17

independently reproduced

Scored on 8 benchmarks. 75% of its scores are independently reproduced — the rest are self-reported or unverified.

8
Benchmarks scored
0
SOTA leads
93.33
Best score
75%
Reproduced
6 of 8

Benchmark fingerprint

Every score, ranked, colored by evidence status.

  • OTIS Mock AIME 2024-2025
    93.33
  • GPQA diamond
    85.1
  • WeirdML
    49.89
  • FrontierMath-Tiers-1-3-v2-Private
    42.81
  • SimpleQA Verified
    38
  • Chess Puzzles
    25
  • FrontierMath-Tier-4-v2-Private
    14.63
  • CritPt
    8

All benchmark scores

BenchmarkScoreEvidence statusMeasured
OTIS Mock AIME 2024-202593.33accuracy (%)reproduced· optimized2026-04-17
GPQA diamond85.1accuracy (%)reproduced· optimized2026-04-17
WeirdML49.89accuracy (%)unverified2026-04-17
FrontierMath-Tiers-1-3-v2-Private42.81accuracy (%)reproduced2026-04-17
SimpleQA Verified38accuracy (%)reproduced2026-04-17
Chess Puzzles25accuracy (%)reproduced2026-04-17
FrontierMath-Tier-4-v2-Private14.63accuracy (%)reproduced2026-04-17
CritPt8accuracy (%)unverified2026-04-17
8 source records · epoch_benchmarks. Every score is evidence-linked.

Follow the record

More from xAI

  • Grok 494.4
  • Grok 4 Fast94.4
  • Grok-3 mini90.94
  • Grok 4.2089.5
  • Grok 388.75
  • Grok-2 (Dec 2024)63.6