Benchmark

RefSpatialBench

RefSpatialBench evaluates spatial reference understanding and grounding.

Modality
image
Categories
spatial_reasoning, grounding, vision
Openness
unknown
Source
llm_stats
Reported scores
6

Reported scores

llm_stats

ModelOrganizationReported valueReportedEvidence
Qwen3 VL 235B A22B ThinkingQwen0.6992025-09-22evidence
Qwen3.5-122B-A10BQwen0.6932026-02-24evidence
Qwen3.5-27BQwen0.6772026-02-24evidence
Qwen3.5-35B-A3BQwen0.6352026-02-24evidence
Qwen3.6-27BQwen0.72026-04-21evidence
Qwen3.6-35B-A3BQwen0.6432026-04-16evidence

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.