AA-Briefcase Elo
Agents · 2026-06-18
AA-Briefcase is an Artificial Analysis benchmark for long-horizon agentic knowledge-work projects. It evaluates agents on realistic multi-week business workflows with many linked tasks and thousands of input files, combining rubric and pairwise grading into an Elo-style score.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Opus 5 | 1720 |
| Claude Fable 5.1 | 1662 |
| Qwen3.8 Max (0902) | 1622 |
| Muse Spark 1.3 | 1589 |
| Qwen3.8-Flash-Next | 1587 |
| GPT-6 Astra | 1562 |
| Kimi K3 | 1548 |
| Grok 4.6 | 1545 |
| Claude Fable 5 | 1530 |
| GLM-5.3 | 1505 |
| GPT-5.6 Sol | 1495 |
| GLM-5.3 Flash | 1449 |
| Qwen3.8 2.4T A95B | 1439 |
| DeepSeek V4 Flash Vision Exp | 1432 |
| DeepSeek V4.1 Flash | 1424 |
| Qwen3.8 27B | 1398 |
| Qwen3.8-Max | 1388 |
| Sonnet 5 | 1355 |
| Opus 4.8 | 1354 |
| Muse Spark 1.2 | 1344 |
| GPT-5.6 Terra | 1337 |
| GPT-5.6 Luna | 1333 |
| K2 Horizon 375B A23B | 1298 |
| Grok 4.5 | 1283 |
| DeepSeek V4 Pro 0813 | 1265 |
| GLM-5.2 | 1260 |
| DeepSeek-V4-Flash-0731 | 1253 |
| Opus 4.7 | 1252 |
| Gemini 3.8 Flash | 1202 |
| Agnes 2.5 Pro Alpha | 1195 |
| K2 Horizon MoVA 36B A4B | 1170 |
| GPT-5.5 | 1158 |
| Quasar 438B | 1123 |
| Gemini 3.7 Flash | 1116 |
| MiniMax M3 | 1096 |
| GPT-5.5 Instant (2026-06-25 hosted snapshot) | 1082 |
| Sonnet 4.6 | 1065 |
| DeepSeek-V4-Flash | 1052 |
| K2 Horizon 7B | 1021 |
| Ling 3.0 Flash VL | 986 |