FinanceQA - Basic Tactical
Finance & Accounting · 2025-01-29
The basic tactical FinanceQA component: 38 questions that ask for a figure or ratio computed from the supplied financial statements and notes following professional accounting and valuation conventions, with every needed input present. Answers are graded as exact correct matches.
Top models (higher is better)
| Model | Score |
|---|---|
| o3 | 55.3 |
| Gemini 2.5 Pro | 55.3 |
| Grok 4 | 52.6 |
| O4 Mini | 52.6 |
| Sonnet 4 | 52.6 |
| DeepSeek-R1-0528 | 50.0 |
| Grok 3 (API) | 47.4 |
| Opus 4 | 44.7 |
| QwQ-32B | 44.7 |
| Qwen3-30B-A3B | 42.1 |
| Llama 4 Maverick Instruct | 39.5 |
| Phi-4-reasoning-plus | 39.5 |
| GPT-4.1 Mini | 39.5 |
| Kimi K2 Instruct | 39.5 |
| Command A | 39.5 |
| Llama 3.1 Nemotron Ultra 253B V1 | 36.8 |
| Magistral Medium 1.0 | 34.2 |
| Gemini 2.5 Flash | 31.6 |
| Nova Pro | 28.9 |