Skip to content

Models Benchmarks Providers

Search models and benchmarks /

ZEROBench

reasoning

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

ZEROBench is a challenging vision benchmark designed to test models on zero-shot visual understanding tasks.

Leaderboard

Showing 6 of 6 results

Muse Spark

33.0%

i
Kimi K2.5

11.0%

i
Qwen3.5-27B

10.0%

i
Qwen3.5-122B-A10B

9.0%

i
Qwen3.5-35B-A3B

8.0%

i
Qwen3 VL 235B A22B Thinking

4.0%

i

Wikibench About Theme Content licensed CC BY-SA 4.0.