Qwen: Qwen3 VL 235B A22B Thinking

Qwen3-VL-235B-A22B Thinking is a cutting-edge multimodal model designed for advanced reasoning, particularly in STEM and mathematics. It seamlessly integrates robust text generation with sophisticated visual understanding capabilities, processing both images and video. This model is optimized for multimodal reasoning, demonstrating competitive results on public benchmarks for perception and reasoning, including robust recognition of diverse real-world and synthetic categories, spatial understanding (2D/3D grounding), and long-form visual comprehension. Beyond analytical tasks, Qwen3-VL supports agentic interaction and tool use. It can follow complex instructions in multi-image, multi-turn dialogues, align text to video timelines for precise temporal queries, and operate GUI elements for automation. The model also facilitates visual coding workflows, converting sketches or mockups into code and assisting with UI debugging. It maintains strong text-only performance comparable to flagship Qwen3 language models, making it suitable for production scenarios such as document AI, multilingual OCR, software/UI assistance, spatial/embodied tasks, and vision-language agent research. Key specifications include a large context window of 262K tokens and a max output of 4K tokens. It offers capabilities like vision, functions, code, and streaming. Pricing is competitive at $0.45 per 1M input tokens and $3.50 per 1M output tokens. Access is available under the PRO tier.