SnorkelFinance
Professional Work · 2026-06-08
SnorkelFinance is an expert-verified financial question-answering benchmark from Snorkel AI. It evaluates ReAct-style agents on 290 financial QA tasks derived from real reports using SQL, code, and other tools, with scores reported as overall accuracy from full execution traces.
Top models (higher is better)
| Model | Score |
|---|---|
| GPT-5 | 81.0 |
| o3 | 81.0 |
| Gemini 3 Pro Preview | 80.3 |
| Opus 4.1 | 80.3 |
| GPT-5 Mini | 79.3 |
| Opus 4 | 78.3 |
| Claude 3.7 Sonnet | 77.9 |
| Sonnet 4 | 76.6 |
| O4 Mini | 76.6 |
| Grok 4 | 74.0 |
| Grok 4 Fast | 73.5 |
| Kimi K2 Thinking | 71.7 |
| gpt-oss-120b | 66.6 |
| Grok 3 | 65.9 |
| o3-mini | 63.8 |
| GPT-4.1 | 62.7 |
| Nova Premier | 62.1 |
| Gemini 2.5 Pro | 60.6 |
| Gemini 2.5 Flash | 53.1 |
| Qwen3-235B-A22B | 51.4 |
| GPT-5 Nano | 50.0 |
| Llama 3.3 Nemotron Super 49B V1.5 | 44.0 |
| Nova Pro | 40.3 |
| Codestral 25.01 | 27.6 |
| Nova Lite | 16.9 |
| Magistral Medium 1.0 | 16.2 |
| Nova Micro | 14.5 |
| Mistral Large 1.0 | 13.4 |