Cohere launches Transcribe, an open-source ASR model that tops accuracy benchmarks

Cohere launches an open-source voice model specifically for transcription | TechCrunch

Enterprise AI company Cohere on Thursday launched its first voice model: Transcribe is an open-source automatic speech recognition model that can be used for tasks like note-taking and speech analysis. Relatively light at just 2 billion parameters, the model is meant for use with consumer-grade GPUs for those who want to self-host it. It currently supports 14 languages: English, French, German, Italian, Spanish, Portuguese, Greek, Dutch, Polish, Chinese, Japanese, Korean, Vietnamese and Arabic. Cohere says Transcribe beats models such as Zoom Scribe v1, IBM Granite 4.0 1B, ElevenLabs Scribe v2, and Qwen3-ASR-1.7B Speech on the Hugging Face Open ASR leaderboard, achieving an average word error rate (WER) of 5.42, lower than any other model on the benchmark. The company claims Transcribe had an average win rate of 61% over other models when human evaluators assessed its transcriptions for accuracy, coherence and usability. However, the model fell behind its rivals when it had to transcribe Portuguese, German and Spanish. Cohere says Transcribe can process 525 minutes of audio in a minute, which is high for its class of model. The company is planning to integrate Transcribe into its enterprise agent orchestration platform, North, and is making the model available through its API for free. The model will also be available on Model Valut, Cohere's managed inference platform. Speech recognition models are growing increasingly popular as demand grows for note-taking and dictation apps like Granola and Wispr Flow. Earlier this year, Cohere reportedly told investors that it was generating annual recurring revenue of $240 million in 2025, and its CEO, Aidan Gomez, was cited as saying that the startup may go public "soon".

Cohere Launches Transcribe: What's Next?

Cohere Expands Into Audio Market With New Transcription Capabilities Cohere has launched Transcribe, an automatic speech recognition (ASR) model, now available for download as open source. The model, designed for real-world application, is intended to enhance AI-driven tasks such as meeting transcription and speech analytics to support customer support interactions, a press release stated. "Our objective was straightforward: push the frontier of dedicated ASR model accuracy under practical conditions. The model was trained from scratch with a deliberate focus on minimizing word error rate (WER), while keeping production readiness top-of-mind. In other words, not just a research artifact, but a system designed for everyday use," Cohere stated. What Does Transcribe Do? Transcribe supports 14 languages, including English, French and Chinese. The model ranks first for accuracy on Hugging Face's Open ASR Leaderboard, outperforming other ASR models such as Whisper Large v3 and ElevenLabs Scribe v2, with a WER of just 5.42%. In the future, Cohere plans to integrate Transcribe more deeply with its AI agent orchestration platform, North, aiming to expand its capabilities beyond transcription. The model is available for download on Hugging Face and users can access it via Cohere's API for experimentation or through the Model Vault for production deployment. What's Next Earlier this week, Cohere announced a strategic partnership with Saab to advance artificial intelligence technologies. Through this agreement, Saab and Cohere aim to leverage their combined expertise to support high-value industrial cooperation in Canada, marking a significant step forward in AI integration within the aerospace sector. Cohere was founded in 2019 in Toronto and specializes in Large Language Models (LLMs) and generative AI with a specific focus on serving the enterprise market. The company builds AI technology tailored for businesses, focusing on security, data privacy and customizability. Photo: Shutterstock This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors. Market News and Data brought to you by Benzinga APIs To add Benzinga News as your preferred source on Google, click here.