All Benchmarks/Factuality

FACTS Search

Higher is better

Questions about obscure subjects or requiring several hops, which the model must answer using a web search tool. Scores run from 0 to 100%. Higher is better.

Other versions:FACTS Benchmark SuiteFACTS GroundingFACTS MultimodalFACTS Parametric
Top score89.9%GPT-6 Astra
Models tested17
Last updated2026.09.29
Model release date (newest on the right)
Best score so farHigher on the chart is better
  1. 1
    OpenAI
    GPT-6 Astra
    2026.09.04
    89.9%
  2. 2
    OpenAI
    GPT-5.6 Sol
    2026.07.09
    88.5%
  3. 3
    Anthropic
    Claude Opus 5.5
    2026.09.22
    86.5%
  4. 4
    OpenAI
    GPT-6 Sol
    2026.09.22
    86.1%
  5. 5
    Google
    Gemini 3.1 Pro
    2026.02.19
    83.7%
  6. 6
    OpenAI
    GPT-5.6 Terra
    2026.07.09
    81.7%
  7. 7
    OpenAI
    GPT-6 Luna
    2026.09.22
    81.4%
  8. 8
    Google
    Gemini 3 Flash
    2025.12.17
    80.9%
  9. 9
    Google
    Gemini 3.7 Flash
    2026.08.13
    80.9%
  10. 10
    Anthropic
    Claude Opus 4.8
    2026.05.27
    79.6%