Agentic benchmark
Agents' Last Exam leaderboard
Every model the catalog carries a published Agents' Last Exam value for, ranked by that value.
CategoryAgentic
MeasureProvider-reported task score
TasksAgent tasks
DifficultyAdvanced agentic work
Published byDeepSeek V4 Flash 0731 update
An agent benchmark reported in DeepSeek's V4 Flash 0731 launch comparison.
Agents' Last Exam ranking
11 models with a published Agents' Last Exam value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.
| Rank | Model | Provider | Provider-reported task score |
|---|---|---|---|
| 1 |  GPT-6 Astra | OpenAI | 59.3 |
| 2 |  Qwen3.8 Max | Alibaba | 52.4 |
| 3 | Qwen3.8-Flash-Next | Alibaba | 51.2 |
| 4 | Qwen3.8-27B | Alibaba | 42.9 |
| 5 |  DeepSeek V4.1 Flash | DeepSeek | 31.8 |
| 6 |  GLM-5.3 | Z.AI | 28.5 |
| 7 |  Gemini 3.7 Flash | 26.3 | |
| 7 | GLM-5.3-Flash | Z.AI | 26.3 |
| 9 | DeepSeek V4 Pro 0813 | DeepSeek | 25.7 |
| 10 | DeepSeek V4 Flash 0731 | DeepSeek | 25.2 |
| 11 |  Hy4 preview | Tencent | 22.8 |
Evidence key: Observed
Rows are ordered by the value DeepSeek V4 Flash 0731 update published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.