Atlas

Benchmarks

← All benchmarks

WorldVQA Objects F-score

Multimodal · 2026-01-28

WorldVQA category F-score on 437 Objects & Products questions.

Top models (higher is better)

ModelScore
Gemini 3 Pro Preview48.1
Kimi K2.544.7
Gemini 2.5 Pro39.6
Opus 4.539.6
Seed1.5-VL32.8
GPT-5.226.6
GPT-5.126.6
Qwen3-VL-235B-A22B-Instruct26.1
Grok 4.1 Fast25.2
Sonnet 4.522.9
Grok 4 Fast22.0
GLM-4.6V19.2
GPT-4o19.1
Qwen3 VL 32B Instruct19.0
GLM 4.6V Flash14.9
Loading Atlas data…