Omni2Web - Direct Editing Revision Handling: leaderboard

Metric: Revision-handling rate (%; Track A rubric items: revision and undo steps are honoured in the final page; 918 bilingual screen recordings of weakly referential webpage edit requests (speech plus cursor), 13,907 edit steps; the model edits the source HTML directly from the recording under its native input interface (omni: audio-visual video; vision-language: video or 1 fps frames plus manual transcript); rubric items judged by Gemini 3.1 Pro; unusable outputs score zero). Source: arxiv.org. Saturation forecast: Around August 2027. 15 models tracked.

Top models

#ModelScore
1Gemini 3.5 Flash71.26
2GPT-5.6 Sol63.49
3Claude Opus 561.67
4Seed 2.1 Pro57.03
5Kimi K356.59
6Qwen 3.8 Max55.77
7Grok 4.651.43
8MiniMax-M350.87
9Gemini 3.1 Pro (Preview)41.67
10MiMo-V2.539.81
11Seed 2.0 Lite38.87
12Muse Spark 1.137.91
13Qwen 3.5 Omni Plus32.28
14Qwen 3.8 27B26.74

Interactive version: theaggregate.ai/benchmark?slug=omni2web-direct-editing-revision-handling · How It Works · Data refreshed daily, snapshot 2026-09-26.