Professional Reasoning Benchmark - Finance
Professional Work · 2025-11-13
Weighted score on finance reasoning tasks drawn from real professional practice.
Top models (higher is better)
| Model | Score |
|---|---|
| Muse Spark 1.3 | 59.5 |
| Muse Wonder | 58.7 |
| Muse Spark 1.1 | 55.0 |
| Claude Fable 5 | 53.9 |
| Opus 4.6 | 53.3 |
| Muse Spark | 52.4 |
| GPT-5 | 51.3 |
| GPT-5 Pro | 51.1 |
| Claude Fable 5.1 | 50.8 |
| GPT-5.6 Sol | 50.5 |
| Gemini 4 Argon | 49.9 |
| Gemini 3.8 Flash | 49.5 |
| o3-pro | 49.1 |
| GPT-5.1 | 48.0 |
| o3 | 47.7 |
| GPT-6 Astra | 47.5 |
| DeepSeek V4.1 Flash 1017 | 47.1 |
| Kimi K2.5 | 46.5 |
| GPT-5.2 Pro | 46.3 |
| GPT-6.1 Sol | 46.2 |
| Opus 4.5 | 46.2 |
| GPT-5.4 | 45.6 |
| gpt-oss-120b | 43.8 |
| Sonnet 4.5 | 43.8 |
| Kimi K2 Thinking | 43.4 |
| Gemini 3.1 Pro Preview | 41.9 |
| O4 Mini | 39.2 |
| Gemini 3 Pro Preview | 39.2 |
| Qwen3 235B A22B Instruct 2507 | 39.1 |
| Gemini 2.5 Pro | 38.9 |
| Gemini 2.5 Flash | 38.4 |
| Kimi K2 Instruct | 38.3 |
| Opus 4.1 | 35.2 |
| DeepSeek-V3.1 | 35.1 |
| GPT-4.1 | 34.3 |
| DeepSeek-R1-0528 | 32.7 |
| GPT-4.1 Mini | 30.4 |
| Llama 4 Maverick Instruct | 22.4 |