Model overview
What this model is for
Qwen: Qwen3 VL 30B A3B Instruct is a multimodal model for text generation and visual understanding, with listed text and image modalities and advertised support for images and videos. Its Instruct variant is designed for instruction-following across general multimodal tasks, making it suitable for image analysis, visual question answering, document or screen interpretation, and applications that combine textual prompts with visual inputs. Tool-calling is listed as a capability, so it may also fit workflows where model responses need to invoke external functions, although implementation details and supported tool schemas are not available in the supplied material. The advertised context window is 262,144 tokens; no separate runtime default or num_ctx setting is provided. OpenRouter lists pricing of $0.15 per million input tokens and $0.60 per million output tokens. The record identifies approximately 30 billion parameters, but does not provide an inference-speed benchmark, evaluation results, or deployment requirements. Licensing terms, safety behavior, video-input implementation details, and provider-specific limits are not available in the supplied material, so they should be verified before production use.