UI-Venus-2 Report - DeskCraft (Standard + Interactive): leaderboard
DeskCraft targets long-horizon professional desktop workflows and proactive human-agent collaboration such as clarification requests and interruptions. The UI-Venus-2 technical report gives an author-evaluated aggregate over the 538-task union of the Standard and Interactive splits, not the official split-level results; percent, higher is better.
Metric: Success Rate (%). Source: arxiv.org. Saturation forecast: Around February 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Kimi K2.6 | 41.4 | #113 |
| 2 | Seed 2.0 Pro | 40 | #122 |
| 3 | Qwen 3.6 27B | 28.7 | #182 |
| 4 | Qwen 3.5 9B | 14.6 | #391 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=ui-venus-2-report-deskcraft-standard-plus-interactive · How It Works · Data refreshed daily, snapshot 2026-11-05.