Wikibench

ZEROBench

reasoning

Multilingual: no

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

Methodology

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

Leaderboard

Showing 6 of 6 results

  1. Muse Spark self-reported llm-stats
    33.0%
  2. Kimi K2.5 self-reported llm-stats
    11.0%
  3. Qwen3.5-27B self-reported llm-stats
    10.0%
  4. Qwen3.5-122B-A10B self-reported llm-stats
    9.0%
  5. Qwen3.5-35B-A3B self-reported llm-stats
    8.0%
  6. Qwen3 VL 235B A22B Thinking self-reported llm-stats
    4.0%