Qwen 3 model with 8B parameters from Alibaba. Multimodal vision-language model from the Qwen3-VL series, built for high-fidelity understanding and reasoning across text, images, and video.