Finance AI Benchmark
Model Leaderboard
Compare frontier models on finance knowledge, statement analysis, and derivative pricing tasks. Scores use Pass@1 across 45 tasks in three categories.
| Model | Provider | Pass@1 | Overall | Knowledge | Analysis | Quant | Completed |
|---|---|---|---|---|---|---|---|
| #1gpt-5.5 | openai | 82.2% | 78.5% | 100.0% | 100.0% | 46.7% | 6/10/2026 |
| #2gpt-5.4 | openai | 80.0% | 78.5% | 100.0% | 100.0% | 40.0% | 6/10/2026 |
| #3gpt-5.4-mini | openai | 75.6% | 71.9% | 100.0% | 100.0% | 26.7% | 6/10/2026 |
| #4claude-sonnet-4-6 | anthropic | 73.3% | 59.3% | 100.0% | 80.0% | 40.0% | 6/10/2026 |
| #5claude-fable-5 | anthropic | 73.3% | 68.9% | 100.0% | 100.0% | 20.0% | 6/10/2026 |
| #6gpt-4o | openai | 71.1% | 67.4% | 100.0% | 93.3% | 20.0% | 6/10/2026 |
| #7claude-haiku-4-5-20251001 | anthropic | 68.9% | 55.6% | 100.0% | 80.0% | 26.7% | 6/10/2026 |
| #8claude-opus-4-8 | anthropic | 62.2% | 50.4% | 100.0% | 73.3% | 13.3% | 6/10/2026 |