AA-Briefcase v1.1 Elo
Agents
AA-Briefcase v1.1 evaluates 91 tasks across four multi-week professional projects. Its Elo combines rubric completion, analytical quality, and presentation quality. Tasks are run independently, without carrying over a model's earlier submissions. This protocol version is recorded separately from the original AA-Briefcase.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Sonnet 5.5 | 1812 |
| Claude Opus 5.5 | 1800 |
| Muse Wonder Full | 1729 |
| Kimi K3.1 Pro | 1678 |
| Claude Fable 5.1 | 1669 |
| Muse Wonder | 1659 |
| Claude Opus 5 | 1655 |
| Grok 4.7 | 1638 |
| Gemini 4 Carbon | 1609 |
| Qwen3.8 Max (0902) | 1608 |
| Muse Spark 1.3 | 1581 |
| Claude Haiku 5.5 | 1575 |
| Qwen3.8-Flash-Next | 1574 |
| DeepSeek V4.1 Flash 1017 | 1570 |
| GPT-6 Astra | 1568 |
| GPT-6.1 Sol | 1555 |
| Grok 4.6 | 1541 |
| Claude Fable 5 | 1535 |
| Ling 3.1 Flash 1019 | 1521 |
| MiMo-V2.6-Pro | 1511 |
| GLM-5.3 | 1504 |
| Kimi K3 | 1498 |
| MiMo-V2.6-Flash | 1486 |
| Gemini 4 Argon | 1483 |
| GPT-5.6 Sol | 1478 |
| GPT-6 Sol | 1476 |
| MiniMax M3.1 Flash | 1473 |
| GLM-5.3 Flash | 1450 |
| Qwen3.8 2.4T A95B | 1431 |
| Step 5 Preview | 1423 |
| DeepSeek V4 Flash Vision Exp | 1420 |
| DeepSeek V4.1 Flash | 1417 |
| GPT-6.1 Terra | 1402 |
| Ling 3.1 Flash | 1398 |
| Qwen3.8 27B | 1393 |
| Mistral Large 4 Preview | 1391 |
| Qwen3.8-Max | 1382 |
| Sonnet 5 | 1356 |
| GPT-5.6 Luna | 1339 |
| GPT-6 Luna | 1336 |