AssemblyAI Launches Universal-3.5 Pro Speech-to-Text Model on OpenRouter

AssemblyAI's Universal-3.5 Pro speech-to-text model is now available on OpenRouter. The model can transcribe audio in 19 languages in a single synchronous call and ranks first in accuracy across multiple independent speech-to-text benchmarks.

According to the announcement, Universal-3.5 Pro accepts audio clips up to two minutes long and returns the full transcript with word-level timestamps and confidence scores. Developers can also steer transcription toward their domain by supplying keyterms and contextual prompts.

AssemblyAI is a company focused on speech AI technology, offering speech-to-text and audio intelligence APIs. OpenRouter is a platform that aggregates AI models from multiple providers, allowing developers to access models like Universal-3.5 Pro through a unified API.

The model is now live on OpenRouter, with a 50% discount available through September 29, 2026.