Benchmark

V*

A visual reasoning benchmark evaluating multimodal inference under challenging spatial and grounded tasks.

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

Reported scores

llm_stats

ModelOrganizationReported valueReportedEvidence
GLM-5V-TurboZ.ai0.892026-04-02evidence
Kimi K2.6Moonshot AI0.9692026-04-20evidence
Qwen3.5-122B-A10BQwen0.9322026-02-24evidence
Qwen3.5-27BQwen0.9372026-02-24evidence
Qwen3.5-35B-A3BQwen0.9272026-02-24evidence
Qwen3.6 PlusQwen0.9692026-04-02evidence
Qwen3.6-27BQwen0.9472026-04-21evidence

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.