ZeroEval GPQA Diamond — leaderboard
Allen AI ZeroEval evaluation of GPQA Diamond: graduate-level science questions requiring PhD-level reasoning, scored independently from Artificial Analysis.
Metric: GPQA Diamond Score. Source: www.vals.ai. Status: saturated. 232 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.6 Sol | 94.6 |
| 2 | Claude Mythos Preview | 94.6 |
| 3 | Gemini 3.1 Pro (Preview) | 94.3 |
| 4 | Claude Opus 4.7 | 94.2 |
| 5 | GPT-5.5 | 93.6 |
| 6 | Claude Opus 4.8 | 93.6 |
| 7 | Kimi K3 | 93.5 |
| 8 | GPT-5.2 Pro | 93.2 |
| 9 | Grok 4.5 | 93 |
| 10 | GPT-5.6 Terra | 92.9 |
| 11 | GPT-5.4 | 92.8 |
| 12 | GPT-5.2 | 92.4 |
| 13 | Qwen 3.7 Max | 92.4 |
| 14 | GPT-5.6 Luna | 92.3 |
| 15 | Gemini 3 Pro | 91.9 |
Interactive version: theaggregate.ai/benchmark?slug=zeroeval-gpqa-diamond · How the rankings work · Data refreshed daily, snapshot 2026-07-22.