Cohere Releases Open-Weight ASR Model 'Transcribe': 5.4% WER Beats Whisper and ElevenLabs

2026-03-31T12:53:52.553Z·2 min read
Cohere has released Transcribe, a 2-billion parameter open-weight automatic speech recognition (ASR) model that achieves a 5.42% word error rate (WER), outperforming industry leaders including Open...

Cohere has released Transcribe, a 2-billion parameter open-weight automatic speech recognition (ASR) model that achieves a 5.42% word error rate (WER), outperforming industry leaders including OpenAI's Whisper and ElevenLabs' Scribe.

Performance Benchmarks

ModelWERLicense
Cohere Transcribe5.42%Apache-2.0
ElevenLabs Scribe v25.83%Proprietary
Qwen3-ASR-1.7B5.76%Open-weight
Whisper Large v37.44%MIT

Transcribe currently tops the Hugging Face ASR leaderboard.

Key Features

Why This Matters

Enterprise transcription has been a trade-off:

Transcribe breaks this tradeoff by delivering best-in-class accuracy with the ability to run on-premises.

Use Cases

Analysis

Cohere's Transcribe is strategically significant for several reasons. First, it demonstrates that open-weight models can match or exceed proprietary alternatives — a trend we're seeing across AI modalities. Second, the Apache-2.0 license means any company can deploy it without vendor lock-in. Third, by running on local infrastructure, it solves the data residency and latency concerns that prevent many enterprises from using cloud-based transcription APIs.

For teams building AI agent pipelines with voice interfaces, Transcribe provides a production-ready path to transcription without the compromises of closed APIs.

← Previous: UBTech Revenue Surges 53% to 2 Billion Yuan as Humanoid Robots Become Top Revenue SourceNext: Midjourney Engineer Releases Pretext: A 15KB Open-Source Library That Revolutionizes Web Text Layout →
Comments0