Runway launches Media Router to auto-select AI models for images, video, and audio

Runway released Media Router, a tool that automatically selects the most appropriate AI model for generating images, videos, and audio based on specified criteria. Rather than forcing creators to manually choose among a growing thicket of generative models, Media Router routes each request to the best-suited engine — abstracting model selection the way an LLM router abstracts text-model choice.

The launch reflects a maturing generative-media market where the proliferation of specialized models (image, video, audio, each with speed/quality/cost tradeoffs) has itself become a usability problem. Media Router positions Runway as an orchestration layer, echoing broader industry moves toward model-routing and 'right model for the task' philosophies articulated this week by Microsoft's Nadella.

Strategically, it deepens Runway's role as a creative platform rather than a single-model provider, potentially incorporating third-party and open models alongside its own. It competes for the professional media-production workflow against NVIDIA's Cosmos generation stack and other creative-AI suites.

The open question is transparency and control — creators may want to know which model produced a given asset and retain override ability, and routing quality depends on how well Runway's selection criteria map to real creative intent. Watch for details on which models are in the router, pricing, and whether creators embrace automated selection over manual control.