EvoGenUI-Bench - Episode Pass: leaderboard

Metric: Episode pass rate (%; mean of the three suites' shares of five-turn episodes in which all five turns pass; 150 five-turn generative-UI tasks (750 turns) in three suites of 50: information presentation, executable interaction and tool-grounded external state; each turn's artifact is executed in a browser and a MiMo-V2.5 evaluator checks hidden requirements from screenshots, source and DOM evidence, MiMo-V2.5 actor traces and runtime logs; turns left unexecuted after an earlier build failure or invalid output count as failures; temperature 0). Source: arxiv.org. Saturation forecast: Around December 2027. 8 models tracked.

Top models

#ModelScore
1Claude Opus 4.737.3
2GPT-5.521.3
3Claude Haiku 4.5 (20251001)13.3
4Qwen 3.6 Plus7.3
5Gemini 3.1 Pro (Preview)6.7
6Gemini 3 Flash (Preview)4.7
7Qwen 3 Coder 480B A35B Instruct2
8GLM 4.5 Air1.3

Interactive version: theaggregate.ai/benchmark?slug=evogenui-bench-episode-pass · How It Works · Data refreshed daily, snapshot 2026-09-29.