Wikibench

Hypersim

vision

Multilingual: no

Hypersim evaluates 3D grounding and depth understanding in synthetic indoor scenes.

Methodology

Imported from llm-stats public benchmark metadata. Modality: image. Max score: 100. Categories: 3d, spatial_reasoning, vision. Language: en. Verified by llm-stats: no.

Leaderboard

Showing 4 of 4 results

  1. Qwen3.5-35B-A3B self-reported llm-stats
    13.1%
  2. Qwen3.5-27B self-reported llm-stats
    13.0%
  3. Qwen3.5-122B-A10B self-reported llm-stats
    12.7%
  4. Qwen3 VL 235B A22B Thinking self-reported llm-stats
    11.0%