AppWorld-UL (Code Agent) - Infeasible Tasks: leaderboard

Metric: Interactive task goal completion (%; share of 102 tasks made infeasible, where the agent must tell the user whose AppWorld unit tests all pass, including a test that the agent asked the simulated user for every piece of hidden information it needed; ReAct code agent that writes Python against the AppWorld app APIs; GPT-5.5 user simulator answering from a fixed knowledge set). Source: arxiv.org. Saturation forecast: Around December 2026. 6 models tracked.

Top models

#ModelScore
1Claude Opus 4.761.8
2GPT-5.5 (Medium)61.8
3Qwen 3.7 Max55.9
4GLM-5.151
5Kimi K2.641.2
6DeepSeek V4 Pro34.3

Interactive version: theaggregate.ai/benchmark?slug=appworld-ul-code-agent-infeasible-tasks · How It Works · Data refreshed daily, snapshot 2026-09-29.