VLAN - (a) Humanoid - Semantic Success: leaderboard

Full1931trajectorysplit; user-conditionedlanguageinstruction. SRlan requiresalluser-specificsemanticconstraints; IAS weightedstructuredsimilaritytocanonicaltarget. Seven releasedVLMs; exactInstructlabelsretained; noactionSRderivedmetricused. Metric: Native score (0–1). Source: https://arxiv.org/html/2611.18346v1

Metric: Native score (0–1). Source: arxiv.org. Saturation forecast: Around 2030. 7 models tracked. Updated 17 Nov 2026.

Top models

#ModelScoreOverall rank
1Llama 3.2 11B Instruct0.558#5226
2Molmo2-8B0.523#1629
3Qwen 2.5 VL 3B Instruct0.483#2648
4Idefics3-8B-Llama30.438#5775
5Qwen 3 VL 8B Instruct0.389#1025
6InternVL3.5-8B0.365#1588
7Qwen 2.5 VL 7B Instruct0.275#1893

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

Interactive version: theaggregate.ai/benchmark?slug=vlan-a-humanoid-semantic-success · How It Works · Data refreshed daily, snapshot 2026-11-17.