FinBen - FNS: leaderboard
Metric: Normalized Score. Source: huggingface.co. 21 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek V3 | 37.72 |
| 2 | plutus-8B-instruct | 34.46 |
| 3 | GPT-4o | 25.5 |
| 4 | O3 Mini | 16.95 |
| 5 | Llama 4 Scout Instruct | 16.9 |
| 6 | DeepSeek-R1-Distill-Qwen-32B-Japanese | 14.45 |
| 7 | Llama 3.1 70B Instruct | 13.61 |
| 8 | finma-7B-full | 11.2 |
| 9 | Qwen 2.5 32B Instruct | 9.71 |
| 10 | Gemma 3 4B (IT) | 0.24 |
| 11 | Gemma 3 27B (IT) | 0.21 |
| 12 | Qwen2-Audio-7B-Instruct | 0 |
| 13 | SALMONN-7B | 0 |
Interactive version: theaggregate.ai/benchmark?slug=finben-fns · How It Works · Data refreshed daily, snapshot 2026-09-05.