AA-AnalystAgent (pass@5)
Professional Work
Fraction of AA-AnalystAgent's 80 quantitative analysis questions answered correctly at least once across five independent attempts.
Top models (higher is better)
| Model | Score |
|---|---|
| GPT-6 Luna | 85.0 |
| Claude Haiku 5.5 | 83.8 |
| Gemini 3.1 Pro Preview | 81.3 |
| DeepSeek V4.1 Flash | 80.0 |
| Inkling-Small | 80.0 |
| MiMo-V2.6-Pro | 80.0 |
| Opus 4.8 | 78.8 |
| Claude Sonnet 5.5 | 78.8 |
| GLM-5.3 Flash | 78.8 |
| GPT-5.5 | 78.8 |
| Inkling | 78.8 |
| Sonnet 5 | 77.5 |
| Gemini 3.7 Flash | 77.5 |
| Gemini 3.5 Flash | 76.3 |
| Grok 4.5 | 76.3 |
| Grok 4.6 | 76.3 |
| Qwen3.8 Max (0902) | 76.3 |
| Claude Opus 5.5 | 75.0 |
| DeepSeek-V4-Pro | 75.0 |
| Qwen3.7-Max (source unspecified) | 75.0 |
| Claude Opus 5 | 73.8 |
| MiniMax M3 | 73.8 |
| Claude Fable 5.1 | 72.5 |
| Opus 4.7 | 72.5 |
| Gemini 3.8 Flash | 72.5 |
| GPT-6.1 Sol | 72.5 |
| Grok 4.7 | 72.5 |
| Kimi K3 | 72.5 |
| Sonnet 4.6 | 71.3 |
| GLM-5.3 | 71.3 |
| Kimi K3.1 Pro | 71.3 |
| Muse Spark 1.3 | 71.3 |
| DeepSeek-V4-Flash | 70.0 |
| GPT-5.6 Sol | 70.0 |
| Step 5 Preview | 70.0 |
| Claude Fable 5 | 68.8 |
| MiMo-V2.5-Pro | 68.8 |
| GPT-6 Astra | 67.5 |
| Ling 3.0 Flash Fin | 67.5 |
| Gemini 4 Argon | 66.3 |