Benchmark

CountBench

CountBench evaluates object counting capabilities in visual understanding.

Modality
image
Categories
reasoning, spatial_reasoning, vision
Openness
unknown
Source
llm_stats
Reported scores
7

Reported scores

llm_stats

ModelOrganizationReported valueReportedEvidence
North Micro Vision InstructCohere0.7252026-08-12evidence
Qwen3 VL 235B A22B ThinkingQwen0.9372025-09-22evidence
Qwen3.5-122B-A10BQwen0.972026-02-24evidence
Qwen3.5-27BQwen0.9782026-02-24evidence
Qwen3.5-35B-A3BQwen0.9782026-02-24evidence
Qwen3.6 PlusQwen0.9762026-04-02evidence
Qwen3.6-27BQwen0.9782026-04-21evidence

Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.