Drift-Bench (State-Oriented, With Clarification) - Premise: leaderboard
Metric: Success rate (%; 200 AgentBench operating-system and database tasks with a premise fault (a false assumption), clarification tools answered by a persona user simulator, mean of 3 trials). Source: arxiv.org. Saturation forecast: Around December 2026. 7 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 235B A22B 2507 Instruct | 68.67 |
| 2 | DeepSeek V3.2 | 67.33 |
| 3 | GPT-5.2 | 65.84 |
| 4 | GPT-OSS-120B | 62.67 |
| 5 | Gemini 2.5 Flash | 58.66 |
| 6 | GLM-4.7 | 53.33 |
| 7 | Llama 4 Maverick | 38.66 |
Interactive version: theaggregate.ai/benchmark?slug=drift-bench-state-oriented-with-clarification-premise · How It Works · Data refreshed daily, snapshot 2026-09-25.