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

DeepSeek-V3.1

DeepSeek·released 2025-08-21

unverified

Scored on 4 benchmarks. 0% of its scores are independently reproduced — the rest are self-reported or unverified.

4
Benchmarks scored
0
SOTA leads
85.17
Best score
0%
Reproduced
0 of 4

Benchmark fingerprint

Every score, ranked, colored by evidence status.

  • Lech Mazur Writing
    85.17
  • Fiction.LiveBench
    52.8
  • WeirdML
    38.36
  • SimpleBench
    28

All benchmark scores

BenchmarkScoreEvidence statusMeasured
Lech Mazur Writing85.17accuracy (%)unverified2025-08-21
Fiction.LiveBench52.8accuracy (%)unverified2025-08-21
WeirdML38.36accuracy (%)unverified2025-08-21
SimpleBench28accuracy (%)unverified2025-08-21
4 source records · epoch_benchmarks. Every score is evidence-linked.

Follow the record

More from DeepSeek

  • DeepSeek-V4-Pro96.66
  • DeepSeek-R1 (May 2025)96.64
  • DeepSeek-V393.73
  • DeepSeek-R193.05
  • DeepSeek-V2 (MoE-236B, May 2024)89.6
  • DeepSeek-V3.287.81