ZeroBench

A multi-step visual reasoning benchmark with pass@5 reporting and optional tool use.

ModelsModels with a result.10
Top resultBest result on this benchmark.50.0%Muse Spark Open
Top-3 spreadPoints from first to third.11.0 pts
YearYear of release.—

Result and price

6 of 10 results

Results

10 results
#ModelResultThe score from the source.
01Muse Spark OpenAbove fitFar above this model's index.50.0%
02GPT-5.441.0%
04Muse Spark33.0%
07Kimi K323.0%