VehicleArena - Multi-Agent Arrival: leaderboard

Metric: Arrival rate (%). Source: arxiv.org. Saturation forecast: Around 2028. 9 models tracked.

Top models

#ModelScoreOverall rank
1Qwen 3.8 Max (xHigh)65.6#28 (Qwen 3.8 Max)
2GLM-5.3 Flash (Max)62.5#62 (GLM-5.3 Flash)
3DeepSeek V4.1 Flash (Max)56.3#35 (DeepSeek V4.1 Flash)
4Kimi K3 (Max)40.6#27 (Kimi K3)
5Qwen 3.8 27B (xHigh)40.6#95 (Qwen 3.8 27B)
6Qwen 3 VL 32B Instruct25#278
7Qwen 3 VL 8B Instruct18.8#406
8GPT-5.6 Sol (High)9.4#18 (GPT-5.6 Sol)

No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #8 Claude Fable 5.1.

Interactive version: theaggregate.ai/benchmark?slug=vehiclearena-multi-agent-arrival · How It Works · Data refreshed daily, snapshot 2026-10-13.