Finance Agent v2 - Precedents
Professional Work · 2026-05-19
The Precedents split of Finance Agent v2. This child benchmark separates a source-reported subtask or subtrack from the parent aggregate so scores at different grains do not share one benchmark_slug.
Top models (higher is better)
| Model | Score |
|---|---|
| Gemini 3.5 Flash | 36.4 |
| Muse Spark 1.1 | 33.8 |
| Gemini 3.6 Flash | 30.8 |
| GPT-5.6 Luna | 29.1 |
| Sonnet 5 | 28.8 |
| Claude Fable 5 | 28.1 |
| GPT-5.6 Sol | 26.2 |
| GPT-5.6 Terra | 26.1 |
| Gemini 3.5 Flash-Lite | 26.0 |
| Claude Opus 5 | 25.9 |
| Opus 4.8 | 25.6 |
| Inkling | 24.8 |
| MiniMax M3 | 23.2 |
| Kimi K3 | 23.1 |
| GPT-5.4 Mini | 22.1 |
| GPT-5.5 | 21.9 |
| Opus 4.7 | 21.5 |
| GPT-5.4 Nano | 20.8 |
| Gemini 3.1 Pro Preview | 20.8 |
| MiMo-V2.5-Pro | 20.6 |
| Qwen3.7-Max | 20.0 |
| Kimi K2.6 | 19.5 |
| GLM-5.2 | 19.4 |
| Grok 4.5 | 19.4 |
| Sonnet 4.6 | 18.9 |
| Gemini 3 Flash Preview | 18.1 |
| Qwen3.7-Plus | 18.1 |
| GLM-5.1 | 17.8 |
| MiMo-V2.5 | 17.2 |
| Qwen3.6 Plus (2026-04-02) | 17.1 |
| DeepSeek-V4-Pro | 15.6 |
| Grok 4.20 | 15.5 |
| Grok 4.3 | 15.3 |
| Mistral Medium 3.5 | 12.9 |
| Nemotron 3 Ultra 550B A55B | 12.8 |
| MiniMax M2.7 | 11.9 |
| Gemini 3.1 Flash-Lite Preview | 11.8 |
| Laguna M.1 | 11.1 |
| Haiku 4.5 | 11.0 |
| Laguna XS.2 | 5.8 |