Finance Agent v2 - General Qualitative Analysis
Professional Work · 2026-05-19
The General Qualitative 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 |
|---|---|
| Muse Spark 1.3 | 83.4 |
| Grok 4.6 | 77.8 |
| Muse Spark 1.2 | 77.6 |
| Gemini 4 Argon | 77.0 |
| Claude Fable 5.1 | 77.0 |
| GLM-5.3 Flash | 76.2 |
| Qwen3.8 Max (0902) | 75.1 |
| MiMo-V2.6-Pro | 75.1 |
| Muse Wonder Full | 74.2 |
| DeepSeek V4.1 Flash 1017 | 74.1 |
| MiMo-V2.6-Flash | 72.5 |
| Muse Spark Open | 72.3 |
| Sonnet 4.6 | 71.6 |
| Grok 4.5 | 71.1 |
| Muse Wonder | 70.8 |
| GLM-5.3 | 70.2 |
| Hy4 Preview | 69.4 |
| Grok 4.7 | 68.7 |
| GLM-5.2 | 68.7 |
| GPT-6 Astra | 68.4 |
| Gemini 3.8 Flash | 68.1 |
| GPT-5.6 Terra | 67.8 |
| DeepSeek V4.1 Flash | 67.4 |
| DeepSeek V4 Pro 0813 | 67.4 |
| GPT-5.6 Luna | 66.6 |
| Opus 4.8 | 66.4 |
| Kimi K3 | 66.4 |
| Muse Spark 1.1 | 66.4 |
| GPT-5.5 | 65.6 |
| Opus 4.7 | 65.4 |
| Mistral Large 4 Preview | 65.4 |
| Gemini 3.7 Flash | 65.3 |
| GPT-5.6 Sol | 64.9 |
| Sonnet 5 | 64.7 |
| Step 5 Preview | 63.8 |
| GLM-5.1 | 63.7 |
| Qwen3.7-Max (source unspecified) | 63.5 |
| Qwen3.8-Max | 63.4 |
| GPT-6.1 Sol | 63.2 |
| Inkling | 62.9 |