Finance Agent v2 - Adjustments
Professional Work · 2026-05-19
The Adjustments 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 | 54.9 |
| Claude Opus 5 | 54.6 |
| Gemini 3.6 Flash | 53.6 |
| GPT-5.6 Luna | 52.0 |
| Kimi K3 | 51.5 |
| Claude Fable 5 | 51.1 |
| GPT-5.6 Terra | 50.4 |
| Muse Spark 1.1 | 49.7 |
| GPT-5.6 Sol | 49.2 |
| Sonnet 4.6 | 48.9 |
| Sonnet 5 | 48.7 |
| Opus 4.8 | 48.4 |
| GLM-5.2 | 46.3 |
| GPT-5.5 | 45.7 |
| Opus 4.7 | 45.4 |
| GPT-5.4 Mini | 45.4 |
| GLM-5.1 | 43.3 |
| MiniMax M3 | 42.3 |
| Gemini 3.5 Flash-Lite | 41.7 |
| DeepSeek-V4-Pro | 41.6 |
| Qwen3.7-Max | 41.2 |
| Grok 4.5 | 38.6 |
| Gemini 3.1 Pro Preview | 38.4 |
| Kimi K2.6 | 37.9 |
| Inkling | 37.8 |
| Gemini 3 Flash Preview | 35.9 |
| Qwen3.6 Plus (2026-04-02) | 35.7 |
| Qwen3.7-Plus | 35.5 |
| Nemotron 3 Ultra 550B A55B | 34.4 |
| MiMo-V2.5-Pro | 34.4 |
| GPT-5.4 Nano | 33.0 |
| Grok 4.3 | 31.9 |
| MiMo-V2.5 | 31.7 |
| Gemini 3.1 Flash-Lite Preview | 31.0 |
| Haiku 4.5 | 30.5 |
| Mistral Medium 3.5 | 27.3 |
| MiniMax M2.7 | 21.0 |
| Grok 4.20 | 17.4 |
| Laguna M.1 | 15.8 |
| Laguna XS.2 | 9.6 |