MobilePA-Bench - Sub-agent Collaboration: leaderboard
Joint routing-and-handoff success when a mobile planning agent decomposes a complex task and delegates work to specialized sub-agents in MobilePA-Bench (MobilePA-Bench, arXiv 2608.23035v2, Table 3). Percent; higher is better.
Metric: Routing-and-handoff joint success (%). Source: arxiv.org. Saturation forecast: Around October 2027. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 77.53 | #64 |
| 2 | Claude Fable 5 | 70.79 | #20 |
| 3 | Gemini 3.6 Flash | 66.29 | #57 |
| 4 | Kimi K3 | 62.92 | #35 |
| 5 | Claude Opus 5 | 62.92 | #16 |
| 6 | GLM-5.2 | 61.8 | #80 |
| 7 | Seed 2.1 Pro | 59.55 | #73 |
| 8 | Qwen 3.8 Max | 53.93 | #34 |
| 9 | GPT-5.5 | 51.69 | #33 |
| 10 | Claude Opus 4.8 | 50.56 | #39 |
| 11 | Qwen 3.7 Max | 50.56 | #85 |
| 12 | GPT-5.6 Sol | 49.44 | #22 |
| 13 | Kimi K2.6 | 43.82 | #113 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mobilepa-bench-sub-agent-collaboration · How It Works · Data refreshed daily, snapshot 2026-11-05.