VLAN - Real World 30 Trajectories - ๐€๐œ๐œ๐’๐‘ โ†‘: leaderboard

Independent evaluation cohort. S4T7=30real-worldtrajectories; language/actionfamilymeans andharmonicAccSR. A4T14=fixed200testtrajectories, 2/scene, seed42, nineVLMs, identicalbudget; notinterchangeablewith1931fullsplit. FullsplitDroneSRact/GRinconsistency held separately. Metric: Native success score (0โ€“1). Source: https://arxiv.org/html/2611.18346v1

Metric: Native success score (0โ€“1). Source: arxiv.org. Saturation forecast: Around April 2027. 7 models tracked. Updated 17 Nov 2026.

Top models

#ModelScoreOverall rank
1InternVL3.5-8B0.302#1588
2Qwen 3 VL 8B Instruct0.283#1025
3Qwen 2.5 VL 7B Instruct0.269#1893
4Molmo2-8B0.193#1629
5Idefics3-8B-Llama30.175#5775
6Llama 3.2 11B Instruct0.165#5226
7Qwen 2.5 VL 3B Instruct0.123#2648

No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.

Interactive version: theaggregate.ai/benchmark?slug=vlan-real-world-30-trajectories ยท How It Works ยท Data refreshed daily, snapshot 2026-11-17.