SGR-Bench — leaderboard
Metric: Item-F1 (self-reported). Source: benchmarklist.com. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 | 66.18 |
| 2 | GLM-5.1 | 65.64 |
| 3 | Claude Opus 4.7 | 61.38 |
| 4 | DeepSeek V4 Pro | 60.98 |
| 5 | Gemini 3.1 Pro (Preview) | 59.06 |
| 6 | Qwen 3.6 Plus | 36.88 |
| 7 | Seed 2.0 Pro | 29.88 |
Interactive version: theaggregate.ai/benchmark?slug=sgr-bench · How the rankings work · Data refreshed daily, snapshot 2026-07-22.