Open FinLLM Reasoning - XBRL-Math — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 26 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Fino1-14B | 87.78 |
| 2 | DeepSeek R1 | 86.67 |
| 3 | DeepSeek R1 Distill Llama 70B | 86.67 |
| 4 | QwQ-32B | 84.44 |
| 5 | DeepSeek R1 Distill Qwen 32B | 84.44 |
| 6 | DeepSeek R1 Distill Qwen 14B | 84.44 |
| 7 | Fino1-8B | 84.44 |
| 8 | Qwen2.5-Math-72B-Instruct | 83.33 |
| 9 | DeepSeek R1 Distill Llama 8B | 81.11 |
| 10 | DeepSeek V3 | 76.67 |
| 11 | O1 | 74.44 |
| 12 | GPT-4.5 | 74.44 |
| 13 | GPT-4o | 72.22 |
| 14 | Llama 3.3 70B Instruct | 70 |
| 15 | Qwen 2.5 72B Instruct | 67.78 |
Interactive version: theaggregate.ai/benchmark?slug=open-finllm-reasoning-xbrl-math · How the rankings work · Data refreshed daily, snapshot 2026-07-22.