Omni2Web - Instruction Recovery Target F1: leaderboard
Metric: Target F1 (%; Track B component: recovered edit targets matched one-to-one to the reference targets; 918 bilingual screen recordings of weakly referential webpage edit requests; the model rewrites the recording into an ordered list of explicit target and action instructions without the source HTML; one-to-one matching against reference instructions, semantic equivalence judged by Gemini 3.1 Pro; unusable outputs score zero). Source: arxiv.org. Saturation forecast: Around 2029. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.5 Omni Plus | 49.99 |
| 2 | Claude Opus 5 | 48.06 |
| 3 | Seed 2.0 Lite | 47.91 |
| 4 | GPT-5.6 Sol | 47.51 |
| 5 | Gemini 3.5 Flash | 47.04 |
| 6 | Gemini 3.1 Pro (Preview) | 42.19 |
| 7 | Kimi K3 | 38.64 |
| 8 | Qwen 3.8 Max | 34.84 |
| 9 | Grok 4.6 | 34.6 |
| 10 | MiMo-V2.5 | 33.95 |
| 11 | Seed 2.1 Pro | 33.21 |
| 12 | MiniMax-M3 | 21.25 |
| 13 | Qwen 3.8 27B | 19.82 |
| 14 | Muse Spark 1.1 | 18.45 |
| 15 | MiniCPM-o-4.5 | 17.87 |
Interactive version: theaggregate.ai/benchmark?slug=omni2web-instruction-recovery-target-f1 · How It Works · Data refreshed daily, snapshot 2026-09-26.