CatalogBench: Reliably publish-ready, factual brief

Top 10 of 35 models · % of products, higher is better

  1. 1 Gemini 4 NeonGoogle 78.6%
  2. 2 GPT-6.1 SolOpenAI 75.0%
  3. 2 GPT-6.2 SolOpenAI 75.0%
  4. 2 GPT-6 AstraOpenAI 75.0%
  5. 5 GPT-6.2 AstraOpenAI 73.2%
  6. 6 Claude Fable 5.2Anthropic 71.4%
  7. 7 Muse WonderMeta 66.1%
  8. 7 GPT-6 SolOpenAI 66.1%
  9. 9 GPT-6.1 TerraOpenAI 64.3%
  10. 10 Gemini 4 HydrogenGoogle 62.5%

Results to 19 Nov 2026 · release 3608133e8331Source: Spring Prompt · CC BY 4.0