ProactiveMobile - Multimodal: leaderboard
Metric: Success rate (%) on the 1,832 multimodal items of ProactiveMobile's test set: from a user profile, device status, world information and behavioral trajectory the model must infer the latent intent and output a recommendation with the executable function sequence, judged against the gold intention by Gemini-2.5-Pro, averaged over the L1-L3 difficulty tiers weighted by items; higher is better. Source: arxiv.org. Saturation forecast: Around August 2028. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | O1 | 12.5 | #177 |
| 2 | Gemini 2.5 Pro | 10.75 | #145 |
| 3 | GPT-5 | 8.08 | #91 |
| 4 | GPT-4o | 4.53 | #333 |
| 5 | Qwen 2.5 VL 7B Instruct | 0.82 | #643 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=proactivemobile-multimodal · How It Works · Data refreshed daily, snapshot 2026-10-11.