NVIDIA opens Nemotron 3.5 Transcribe: one model for transcription in 40 languages, translation, speaker labels and audio Q&A, first among open-weight models on the Open ASR English leaderboard
Transcription, speaker labeling and translation used to take several models; now one open-weight model with a prompt does them, with a reusable training-data pipeline alongside.
// Key points
- Weights are on Hugging Face and ModelScope: it transcribes 40 languages, translates 31 of them to or from English and attributes speakers in 13; recordings up to an hour and up to eight speakers; output is steered with natural-language prompts, with word-level timestamps.
- 5.33% average WER on the Hugging Face Open ASR English leaderboard, first among open-weight models; 4.62% mean WER across seven languages; 2.8% AA-WER on Artificial Analysis, tied for first among open-weight models with about 4B active parameters (30B total) versus 24B for the model it ties.
- NVIDIA also released Granary v2 multilingual data and its curation pipeline; in a medical case study, fine-tuning cut WER on held-out speakers by 45.8% relative to the best prompt. Language support differs by task, audio Q&A is trained in English, and quantized versions are coming.
Builder's takeAI Interview currently stitches ASR to a separate diarization step; one model that returns speaker labels and word-level timestamps would remove a whole layer of alignment code. But the leaderboard results are English: check the model card for which tasks support Chinese, run it on our own recorded Chinese interviews, and confirm the license before deciding to switch.
// Background · from #Open models
Full timeline →- Oct 28 Liquid AI opens LFM3 weights: 1.7B to 2B-class text, vision, audio and omni models that fine-tune on a single GPU
- Oct 27 Cloudflare discloses a GLM-5.3 cross-request leak: 155 responses carried other customers' data, caused by a KV pool smaller than the context window plus an SGLang bug
- Oct 26 Ollama 0.40.11 adds text-to-speech, starting with Phonon, 35% faster than Gradium's own runtime on an M5 Max