GLM ASR

GLM-ASR-2512 is Zhipu AI's (Z.AI) next-generation cloud speech recognition model for real-time conversion of speech into high-quality text. It reports a character error rate of about 0.0717 and maintains strong recognition across multiple scenarios and accents, optimized for noisy environments. It supports Chinese, English and some dialects, targeting daily conversation, meeting minutes, work documents and specialized-terminology transcription. An open-source sibling, GLM-ASR-Nano-2512 (1.5B parameters), accompanies it. Ideal for high-accuracy cloud transcription in production service scenarios.

curl https://api.tunerouter.com/v1/audio/transcriptions \
  -H "Authorization: Bearer $TUNEROUTER_API_KEY" \
  -F file=@sample.mp3 \
  -F model=glm-asr-2512

Model ID: glm-asr-2512 · works with OpenAI SDKs — just point base_url at api.tunerouter.com