Runway API Adds ElevenLabs' Eleven v3 for Expressive Text-to-Speech

Runway has added ElevenLabs’ Eleven v3 model to its developer API, giving Runway Dev users access to expressive text-to-speech generation directly through Runway’s own text-to-speech endpoint, rather than needing a separate ElevenLabs integration.

Details

• What’s new: Eleven v3 is now available via the Runway API’s text-to-speech endpoint, bringing ElevenLabs’ most expressive voice model into Runway’s unified media generation stack
• Expressive delivery: The model supports inline audio tags such as [laughs] and [whispers] written directly into the input script, letting developers direct tone and delivery without separate parameters
• Voice presets: Eleven v3 uses the same preset voice library that was already available for the prior Multilingual v2 model, so existing voice selections carry over
• Pricing: Runway bills the new model at 1 credit per 50 characters of input text, with a 1-credit minimum per request
• Same-day model sunset: In the same July 30, 2026 update, Runway discontinued Gen-3 Alpha Turbo (gen3a_turbo) and Gen-4 Aleph (gen4_aleph) from the API, directing developers to Gen-4.5 or Gen-4 Turbo for video generation and Aleph 2.0 for video editing
• Context: This follows Runway’s July 23, 2026 launch of Runway Media Router, which already listed ElevenLabs as one of the third-party model providers it can route audio requests to — this update gives Eleven v3 specifically a first-class, directly callable endpoint rather than only being reachable through the router

What happened next

• Developers can call Eleven v3 immediately through the existing text-to-speech endpoint on Runway Dev (dev.runwayml.com) using their current API credentials
• Full technical details, including request parameters and audio tag syntax, are published in Runway’s API changelog and developer documentation
• No separate ElevenLabs account or API key is required to use the model through Runway’s endpoint