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.

← Compare

DeepSeek-V3 vs GPT-4o (Nov 2024)

Verdict

DeepSeek-V3 leads 5–1 across 7 shared benchmarks.

DeepSeek-V3 5 · GPT-4o (Nov 2024) 1 · 1 tied · higher isn't always better — see caveats.

Per-benchmark head-to-head

DeepSeek-V3GPT-4o (Nov 2024)
  • 48.4
    Aider polyglot
    18.2
  • 0
    CritPt
    0
  • 3.02
    FrontierMath-2025-02-28-Private
    0.6
  • 42.05
    GPQA diamond
    30.51
  • 64.85
    MATH level 5
    49.77
  • 82.93
    MMLU
    84.13
  • 15.75
    OTIS Mock AIME 2024-2025
    6.16
A higher number is not always a better model. Each score is task performance under a disclosed harness, and some of these benchmarks are saturated or scored under mixed harnesses. Follow any benchmark to its integrity grade before reading a win as decisive.

Follow the record

These models

  • DeepSeek-V393.73
  • GPT-4o (Nov 2024)84.13

DeepSeek-V3 vs …

  • vs PowerMoE-3b
  • vs Claude Fable 5
  • vs GPT-5.5
  • vs Gemini 3.1 Pro
  • vs Llama 3.1-405B