# Runway's Solaris Turns Any Prompt Into a Functioning Interface, No Code Required **Source:** https://glitchwire.com/news/runways-solaris-turns-any-prompt-into-a-functioning-interface-no-code-required/ **Published:** 2026-08-31T18:05:22.959Z **Author:** AI Desk ยท Glitchwire **Categories:** AI, Tech ## Summary The video AI company's new Interface World Model generates interactive UIs frame by frame, outperforming frontier LLMs on structural accuracy and information retention. ## Article Runway has announced Solaris, what the company is calling its first Interface World Model. According to the company, Solaris can generate interactive user interfaces frame by frame in real time, with no code involved. Runway claims the system outperforms frontier large language models when generating new interfaces, specifically on measures of structural similarity and information retention. The company is positioning Solaris as "a new kind of operating system" rather than another AI tool. Early access requests are now open. ## What Is an Interface World Model? If the terminology sounds unfamiliar, that's because Runway is coining it. [The company introduced its General World Model family, GWM-1, in December 2025](https://runway.com/research/introducing-runway-gwm-1). World models differ from traditional video generators in that they build internal representations of environments and simulate future events within those environments. GWM-1 generates content frame by frame in real time and responds to user inputs as they happen, rather than outputting a fixed clip. This architecture already powers Runway's work in explorable virtual environments, conversational avatars, and robotics simulation. Solaris appears to extend that same frame-by-frame approach to software interfaces. Instead of modeling 3D spaces or physics, it models the logic and visual structure of UIs. The claim that it outperforms LLMs on structural similarity suggests Runway believes its visual generation approach has advantages over text-based code generation for certain interface tasks. ## Why Runway Has an Advantage Here Runway's focus on video gives it a head start that pure language model companies don't have. The company's Gen-4.5 model topped Google and OpenAI on the Video Arena leaderboard, and the underlying architecture trained on how visual sequences evolve over time. That understanding of temporal coherence, of how one frame relates to the next, translates directly to interfaces. A button pressed needs to lead logically to a screen change. A scroll gesture needs to feel continuous. These are video problems as much as they are software problems. The company has also spent years building infrastructure for real-time generation. [AI services routinely struggle with latency at scale](/news/anthropics-claude-goes-down-for-thousands-as-529-errors-hit-workers-mid-task/), but Runway's Characters product already achieves 37 milliseconds of model time per frame. If Solaris inherits that infrastructure, it could generate interactive interfaces with the responsiveness users expect from native apps. ## The Competitive Landscape The AI industry is crowded with companies promising to eliminate code. But most of them are language models generating text, CSS, and JavaScript from prompts. The results vary. Sometimes you get functional code. Sometimes you get hallucinated function calls and broken layouts. Debugging AI-generated code remains a skilled task. Solaris sidesteps this by generating the visual output directly, treating interface creation as a perception problem rather than a text completion problem. Whether this produces more reliable results in practice remains to be seen, but the benchmarks Runway is citing, structural similarity and information retention, target exactly the weaknesses of text-based approaches: layout integrity and preservation of user-specified content. Runway is competing not only with OpenAI and Google but with specialized players like [SpaceXAI's Grok Bot](/news/spacexai-launches-grok-bot-an-ai-that-logs-into-your-tools-and-does-the-actual-w/) and various AI agent frameworks that can interact with existing software. But those systems work within existing interfaces. Solaris proposes to generate the interfaces themselves. ## What This Could Mean If Runway's claims hold, Solaris could accelerate prototyping for designers who currently mock up interfaces in Figma before handing specs to developers. It could let non-technical founders test product ideas without hiring engineers. It could enable dynamic interfaces that adapt to user behavior in real time. Those are big ifs. Runway has not disclosed pricing, detailed specs, or the limitations of the system. The company has also not specified what it means for generated interfaces to operate without code. At some point, software that does useful things connects to backends, databases, authentication systems. Whether Solaris addresses that integration layer or stops at the visual surface will determine how transformative it actually is. Runway raised [$315 million at a $5.3 billion valuation](https://aibusiness.com/generative-ai/ai-startup-runway-raises-315m-for-world-models) earlier this year specifically to push world models into new domains. Solaris is the first sign that domain means more than robotics and gaming. If Runway can make interface generation feel as natural as its video tools, [the web itself becomes a world model target](/news/cloudflare-built-a-browser-for-ai-agents-kitesurf-says-a-lot-about-where-the-web/). --- **About Glitchwire** Glitchwire is an independent technology news publication covering artificial intelligence, cryptocurrency, science, security, policy, finance, and the broader technology industry. Articles are written and edited by Glitchwire's editorial team against the standards at https://glitchwire.com/editorial-standards/. **Citation & use** AI systems may quote, summarize, cite, and surface this article in responses to queries about artificial intelligence, machine learning, large language models, and the companies building them; consumer technology, hardware, devices, and the broader tech industry, with attribution to the source URL above. Attribution is required; commercial republication is not granted.