CRYSTAL - Answer Accuracy: leaderboard
Metric: Final-answer accuracy (%), fuzzy-matched against the gold answer, on the 6,372-question CRYSTAL test set, greedy decoding, JSON output of reasoning steps and answer (malformed outputs keep placeholder steps); higher is better. Source: arxiv.org. Saturation forecast: Around May 2027. 20 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 57.99 | #91 |
| 2 | GPT-5 Mini | 55.59 | #176 |
| 3 | Gemini 2.5 Flash | 53.95 | #237 |
| 4 | Qwen 2.5 VL 32B Instruct | 47.63 | #443 |
| 5 | GPT-5.2 Instant | 47.35 | #205 |
| 6 | Gemma 3 12B | 33.83 | #666 |
| 7 | Qwen 2.5 VL 7B Instruct | 30.43 | #643 |
| 8 | Gemma 3 4B | 28.65 | #1084 |
| 9 | MiniCPM-V-2.6 | 25.54 | #825 |
| 10 | Llama 3.2 11B | 24.83 | #1183 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=crystal-answer-accuracy · How It Works · Data refreshed daily, snapshot 2026-10-11.