GPQA Diamond (Opus 4.6 System Card): leaderboard
Graduate-level Google-Proof Q&A benchmark evaluating models on questions that require deep reasoning. This slice reports Anthropic's own runs from the Claude Opus 4.6 system card; small frontier-model pool.
Metric: Accuracy (%). Source: www-cdn.anthropic.com. Saturation forecast: Estimated already saturated. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.2 | 93.2 | #110 |
| 2 | Gemini 3 Pro | 91.9 | #80 |
| 3 | Claude Opus 4.6 (Thinking) | 91.3 | #63 (Claude Opus 4.6) |
| 4 | Claude Opus 4.5 (Thinking) | 87 | #83 (Claude Opus 4.5) |
| 5 | Claude Sonnet 4.5 | 83.4 | #142 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=gpqa-diamond-opus-4-6-system-card · How It Works · Data refreshed daily, snapshot 2026-10-20.