Finance Agent v2 - Disclosure Analysis
Professional Work · 2026-05-19
The Disclosure Analysis 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 |
|---|---|
| Claude Opus 5 | 71.3 |
| Claude Fable 5 | 68.8 |
| Sonnet 4.6 | 65.8 |
| Opus 4.8 | 65.7 |
| Kimi K3 | 65.3 |
| Gemini 3.5 Flash | 63.9 |
| Muse Spark 1.1 | 63.4 |
| GPT-5.6 Sol | 63.3 |
| GPT-5.6 Luna | 62.8 |
| Sonnet 5 | 62.1 |
| Gemini 3.6 Flash | 62.0 |
| GLM-5.2 | 61.6 |
| Opus 4.7 | 61.3 |
| GPT-5.6 Terra | 61.3 |
| MiniMax M3 | 59.8 |
| Qwen3.7-Max | 59.2 |
| GPT-5.5 | 58.9 |
| Grok 4.5 | 56.4 |
| Gemini 3.5 Flash-Lite | 56.1 |
| Inkling | 53.2 |
| MiMo-V2.5-Pro | 53.1 |
| GLM-5.1 | 52.7 |
| DeepSeek-V4-Pro | 52.6 |
| Kimi K2.6 | 52.1 |
| Nemotron 3 Ultra 550B A55B | 51.3 |
| GPT-5.4 Mini | 48.3 |
| Grok 4.3 | 47.7 |
| MiMo-V2.5 | 47.3 |
| Gemini 3.1 Pro Preview | 46.2 |
| Gemini 3 Flash Preview | 45.8 |
| Qwen3.6 Plus (2026-04-02) | 44.5 |
| Qwen3.7-Plus | 44.4 |
| GPT-5.4 Nano | 42.8 |
| Haiku 4.5 | 42.5 |
| Mistral Medium 3.5 | 40.2 |
| Grok 4.20 | 40.0 |
| Gemini 3.1 Flash-Lite Preview | 37.4 |
| MiniMax M2.7 | 35.0 |
| Laguna M.1 | 30.5 |
| Laguna XS.2 | 18.7 |