ModelQwen3-VL 2B Instruct
qwen3-vl:2b-instruct
A multimodal vision-language model in the Qwen3-VL series with 2B parameters, showing high accuracy in visual question answering.
Timeline (last 90 days)
Oct 19
On a 4-core CPU, Qwen3-VL 2B-instruct never wrongly answered 'yes' for things that do not exist in the 128 annotated images, but each new frame takes about 18 seconds to read.
Unverified50%