Benchmark

Vibe-Eval

VIBE-Eval is a hard evaluation suite for measuring progress of multimodal language models, consisting of 269 visual understanding prompts with…

Modality
multimodal
Categories
multimodal, general, vision
Openness
unknown
Source
llm_stats
Reported scores
8

Reported scores

llm_stats

ModelOrganizationReported valueReportedEvidence
Gemini 1.5 FlashGoogle0.4892024-05-01evidence
Gemini 1.5 Flash 8BGoogle0.4092024-03-15evidence
Gemini 1.5 ProGoogle0.5392024-05-01evidence
Gemini 2.0 FlashGoogle0.5632024-12-01evidence
Gemini 2.5 FlashGoogle0.6542025-05-20evidence
Gemini 2.5 Flash-LiteGoogle0.5132025-06-17evidence
Gemini 2.5 ProGoogle0.6562025-05-20evidence
Gemini 2.5 Pro Preview 06-05Google0.6722025-06-05evidence

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.