DeskCraft (Interactive): leaderboard

Metric: Task success rate (%) on the 152-task interactive human-in-the-loop split; higher is better. Source: arxiv.org. Saturation forecast: Around December 2026. 18 models tracked.

Top models

#ModelScore
1GPT-5.427.6
2Kimi K2.625.7
3Kimi K2.524
4Qwen 3.5 35B A3B12.4
5Qwen 3.5 397B A17B11.7
6Qwen 3.6 35B A3B6.9
7Qwen 3 VL 32B6.9
8Qwen 3 VL 235B A22B6.2
9Qwen 3.5 9B4.2
10uitars-1.5-7B0

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