Voxtral Transcribe 2

Voxtral models are frontier open source speech‑understanding systems available in two sizes—a 24 B variant for production‑scale applications and a 3 B variant for local and edge deployments, both released under the Apache 2.0 license. They combine high‑accuracy transcription with native semantic understanding, supporting long‑form context (up to 32 K tokens), built‑in Q&A and structured summarization, automatic language detection across major languages, and direct function‑calling to trigger backend workflows from voice. Retaining the text capabilities of their Mistral Small 3.1 backbone, Voxtral handles audio up to 30 minutes for transcription or 40 minutes for understanding and outperforms leading open source and proprietary models on benchmarks such as LibriSpeech, Mozilla Common Voice, and FLEURS. Accessible via download on Hugging Face, API endpoint, or private on‑premises deployment, Voxtral also offers domain‑specific fine‑tuning and advanced enterprise features.

ElevenLabs has introduced Scribe, an advanced Automatic Speech Recognition (ASR) model designed to deliver highly accurate transcriptions across 99 languages. Scribe is engineered to handle diverse real-world audio scenarios, providing features such as word-level timestamps, speaker diarization, and audio-event tagging. Benchmark tests, including FLEURS and Common Voice, demonstrate Scribe's superior performance over leading models like Gemini 2.0 Flash, Whisper Large V3, and Deepgram Nova-3, achieving the lowest word error rates in languages such as Italian (98.7%) and English (96.7%). Notably, Scribe also significantly reduces errors in languages that have been traditionally underserved, including Serbian, Cantonese, and Malayalam, where other models often exhibit error rates exceeding 40%. Developers can integrate Scribe through ElevenLabs' speech-to-text API, receiving structured JSON transcripts that include detailed annotations.