Skip to content

Models Benchmarks Providers

Search models and benchmarks /

Hypersim

vision

Categories: 3d, spatial reasoning, vision
Modality: image
Language: en
Multilingual: No
Max score: 100
Scoring: %, higher is better
Verified by llm-stats: No

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

Leaderboard

Showing 4 of 4 results

Qwen3.5-35B-A3B

13.1%

i
Qwen3.5-27B

13.0%

i
Qwen3.5-122B-A10B

12.7%

i
Qwen3 VL 235B A22B Thinking

11.0%

i

Wikibench About Theme Content licensed CC BY-SA 4.0.