Drift-Bench (Service-Oriented, No Clarification) - Expression: leaderboard

Metric: Success rate (%; 150 StableToolBench G1 API tasks with an expression fault (ambiguous wording), no clarification tools, mean of 3 trials). Source: arxiv.org. Saturation forecast: Around July 2028. 7 models tracked.

Top models

#ModelScore
1Qwen 3 235B A22B 2507 Instruct75.33
2GLM-4.772.67
3DeepSeek V3.272.44
4Llama 4 Maverick68.22
5Gemini 2.5 Flash64.33
6GPT-5.249.22
7GPT-OSS-120B45.67

Interactive version: theaggregate.ai/benchmark?slug=drift-bench-service-oriented-no-clarification-expression · How It Works · Data refreshed daily, snapshot 2026-09-25.