Z.aiGLM Speech
GLM-ASR-Nano-2512
A 1.5-billion-parameter downloadable speech recognizer.
Practical use
Practical guidance. Include low-volume speech in the evaluation rather than normalizing every recording to a clean studio sample. Review short names and numerical sequences where one missing syllable can change the result.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | Downloadable |
| Model identifiers | glm-asr-nano-2512zai-org/GLM-ASR-Nano-2512 |
| Focus | Mandarin, English, Cantonese and quiet speech.GLM-ASR-Nano-2512 official documentation |
| Licence | MIT.GLM-ASR-Nano-2512 official documentation |
| Access distinction | This Nano checkpoint is named separately from the hosted GLM-ASR-2512 service.GLM-ASR-Nano-2512 official documentation |
Related models
- GLM-ASR-2512
Z.ai · GLM Speech
A hosted recognizer for multilingual audio transcription.
Launch date not confirmedAPI A two-stage speech system with voice cloning and streaming inference.
Launch date not confirmedDownloadable