Share your thoughts, 1 month free Claude Pro on usSee more

Home/Benchmarks

Grounding on RefCOCO + g

63.3IoU

Qwen2-VL-2B

Updated 1mo ago

Evaluation Results

Method	Links
Qwen2-VL-2B 2025.12		63.3
Qwen2-VL-2B 2025.12		51.3

SOTA Paper

Qwen2-VL-2B

RLLaVA: An RL-central Framework for Language and Vision Assistants

Dataset

RefCOCO

Follow for update

@wizwand_team Discord

Related Benchmarks

Math Evaluation on Geometry3K (Accuracy)Counting on CLEVR-Count Search on MAT-Search Coding on MAT-Coding

© 2026 wizwand

Blog Contact Changelog Swarm

Privacy Policy Terms of Service FAQs Swarm Docs