Official benchmarks map / ExtractBench
FINAL-Bench/Darwin-180B-RSI-R3 is ranked first on the ExtractBench leaderboard with 90.29, 0.5% ahead of Qwen/Qwen3.8-Flash-Next. Read live from the Hugging Face API on November 6, 2026.
| # | Model | Organization | Score |
|---|---|---|---|
| 1 | FINAL-Bench/Darwin-180B-RSI-R3 | FINAL-Bench | 90.29 |
| 2 | Qwen/Qwen3.8-Flash-Next | Qwen | 89.88 |
| 3 | Qwen/Qwen3.8-27B | Qwen | 89.75 |
| 4 | nitrai-research/Moderato-V1-Pro | nitrai-research | 88.65 |
| 5 | Qwen/Qwen3.6-35B-A3B | Qwen | 88.11 |
| 6 | deepseek-ai/DeepSeek-V4.1-Flash | deepseek-ai | 87.11 |
| 7 | moonshotai/Kimi-K3 | moonshotai | 83.17 |
| 8 | zai-org/GLM-5.3-Flash | zai-org | 80.75 |
| 9 | datalab-to/lift | datalab-to | 78.43 |
| 10 | numind/NuExtract-2.0-8B | numind | 49.42 |
ExtractBench is a dataset that Hugging Face marks as an official benchmark. Its leaderboard is built automatically from the .eval_results files that model repositories publish, so entries are mostly self-reported by model authors and evaluation settings differ between submissions. A verified tag appears only where Hugging Face marks an entry as verified.
Original board: huggingface.co/datasets/llamaindex/ExtractBench
Part of the Hugging Face official benchmarks map by QuantID. Unofficial. Not affiliated with or endorsed by Hugging Face.