aiaimodel.cards
← All benchmarks

OpenAI-Proof Q&A v0

We haven't written a description for this benchmark yet; see the card notes below for how each maker ran it.

Your workloadRAG over docs · 12K in · 600 out · 50K requests/mo · 30% cached
  1. 1GPT-5.4 Mini · OpenAIbest value
    7.5%

“Best value”: no other model here reports a higher score at a lower cost for your workload. Scores are self-reported by each maker with different settings, so treat close calls as ties.

How each maker reported it

  • GPT-5.4 Mini §6.3, Tables 18–20, p.36

    The appendix reports only Preparedness evaluations. Bio values are without counting refusals; counting refusals and safe completions as successes, GPT-5.4 mini scores 47.0%, 34.24%, 77.77% and 33.11%. Tacit Knowledge is cons@32, CTF pass@12, others pass@1.

Only each model's own maker's number is shown; scores a maker reports for rival models are left out.