GB/T-Bench - Direct Review DMTR@8: leaderboard
Metric: Documents with at least 8 correct diagnoses (%). Source: arxiv.org. 14 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.6 Sol | 54.35 |
| 2 | GPT-5.5 | 54.1 |
| 3 | Claude Fable 5 | 47.83 |
| 4 | GPT-5.4 | 38.73 |
| 5 | Qwen 3.7 Plus | 34.43 |
| 6 | Gemini 3.5 Flash | 30.12 |
| 7 | Claude Opus 4.8 | 23.57 |
| 8 | DeepSeek V4 Pro | 21.93 |
| 9 | GPT-5.3 | 18.44 |
| 10 | Qwen 3 235B A22B 2507 Instruct | 14.96 |
| 11 | DeepSeek V4 Flash | 14.55 |
| 12 | GPT-5.4 Mini | 10.25 |
| 13 | Qwen 2.5 VL 72B Instruct | 8.2 |
| 14 | MiniMax-M2.7 | 6.56 |
Interactive version: theaggregate.ai/benchmark?slug=gb-t-bench-direct-review-dmtr-8 · How It Works · Data refreshed daily, snapshot 2026-09-19.