AI startup Runway is no longer interested in being just another company developing artificial intelligence models; instead, it aims to become the infrastructure layer for generative media. On Thursday, the company launched the Runway Media Router through its developer platform, Runway Dev. This platform, released earlier this month, provides access via an application programming interface (API) to a growing list of third-party models for creating images, video, and audio, alongside Runway's own native models. In an exclusive report by TechCrunch, the company presents the new tool as the first of its kind designed for generative media, unlike model routers that have become common in the world of large language models (LLMs).
How Runway’s Media Router Works
The Media Router is a tool that automatically selects the best image, video, or audio generation model for each request, based on priorities set by the developer—whether they prioritize quality, speed, or cost. "The routing really fits into that overall promise of being the easiest one-stop shop for developers to integrate with any type of generative media model," Anthony Maggio, Runway’s chief product officer, told TechCrunch.
This launch marks another step in Runway’s evolution from an AI video startup to an infrastructure company for other businesses building products with generative media. Through the Runway Dev platform, customers—including companies like Adobe, Cloudflare, ElevenLabs, Expedia, Shutterstock, and Quora—can build media generation capabilities directly into their own products and services using Runway’s API, rather than sending their users to Runway’s own app or website.
Navigating Model Proliferation and Geopolitical Preferences
The launch of the router comes at a time when the number of generative media models has skyrocketed, making the process of evaluating and testing new versions highly complex, difficult, and time-consuming for developers. Through the Runway Dev platform, developers can access the latest media models as soon as they are released to the market. "Most developers are not spending the time to really understand the capabilities of each of these models and where they excel or differ based on various types of outputs across video, image, and audio," Maggio explained. "The unique proposition we’re bringing to the table is all of that intelligence around what the best model is for each different use case, and meshing that with preference you apply around the context of your business."
Maggio noted that Chinese generative media models are becoming increasingly popular. However, many businesses building their own products might not feel comfortable working with models originating from China. Therefore, he explains, developers could potentially set a preference for American model providers—a preference that may become more common as the Trump administration explores bans and sanctions against open-source Chinese AI models.
The Challenge of Token-Based Pricing and Runway's Model Shift
Alongside geopolitical preferences, developers can set a variety of other preferences in the system. Maggio explains that customers are primarily interested in routing the model to account for output quality and token pricing. Token pricing became a central topic in 2026, as enterprises and companies that went all-in on agentic AI experienced high costs and particularly burdensome token bills. In the world of large language models (LLMs), model routing based on cost and token pricing has already become common, so it only makes sense that cost-based routing would expand to the generative media space as well.
Interestingly, the launch of the Media Router comes just weeks after Runway replaced its unlimited subscription plans with token-based pricing—a move that drew criticism from some of its users.
The Intelligence Layer Behind Media Quality Evaluation
When it comes to the quality aspect, Maggio explains that deciding which models provide the best quality for a given task is not as straightforward in generative media as it is with language models. This is precisely where the router's intelligence layer comes into play. This layer is based on the expertise that Runway's in-house creative team has developed in evaluating outputs across all media types—aspects such as how video models handle motion, how image models handle composition, or how voice models process lip syncing.
Runway had already done a large portion of the work on building this intelligence layer for its Agent product—a conversational AI creative partner launched by the company last May, aimed at helping turn text prompts into fully edited multi-shot videos and marketing campaigns. Maggio explains that the Runway Media Router essentially takes the same routing technology developed for the company's internal products and packages it for use by external developers.
Intense Competition and Losing the Lead in Video Leaderboards
Runway's current strategy reflects the highly fragmented and competitive state of the generative media landscape today, and the startup's significant need to expand and pivot to maintain its competitive position in the market. Runway’s last AI video model release—the Gen 4.5 model—was launched last December. At that time, the model topped leaderboards, outperforming similar models from leading companies like Google. In the same month, Runway also released its first world model.
Aside from an upgrade to its video editing model, Aleph 2.0, in May, Runway has not released a new, dedicated frontier video model in many months (TechCrunch reached out to the company asking when the startup plans to release the Gen-5 model).
Today, while Aleph 2.0 ranks among the leading video editing models according to data from analytics firm Artificial Analysis, the company's text-to-video and image-to-video models no longer lead the rankings. In the top 20 spots of these rankings, one can currently find models from powerful market-leading players, including Google and Chinese companies ByteDance and Alibaba. Instead of asking developers to bet on a single model maintaining a technological lead over time, the Media Router is based on the working assumption that the best model will change constantly and frequently. This approach keeps Runway in the game and allows it to continue building and developing at the frontier of technology—if not as the company presenting the newest and best AI model, then as the company providing the best orchestration layer in the market.
Building a Full Stack Solution for the Media World
Anastasis Germanidis, co-founder and co-CEO of Runway, acknowledged in a conversation with TechCrunch that the startup was known for a long time primarily for "that end user piece." However, he explained that to achieve this capability, the company had to build a full stack, including a developer platform, a creative tool suite, and an inference layer underneath it all. He notes that the company is seeing growing interest from enterprises and companies for Runway to be present and operate across every part of this technological stack.
"You need great models underneath, but the orchestration increasingly matters a lot because people are building entire campaigns with those models, or they're building entire finished multi-scene generations out of those models," Germanidis told TechCrunch. "It's something that we increasingly had to build—that intelligence layer that comes on top of the pure pixel models. The router is one way in which the benefits of that come to users."
Or as Maggio put it from a broader perspective: "If you zoom out at the one thing Runway has been doing since 2018, it's that we're deeply focused on research, while building for where we think the space is going at the same time."