KyutaiVoice of Reason
GLM 4 Voice of Reason STITCH 9B
STITCH combines silent text reasoning with spoken output, then uses reinforcement learning. It builds on GLM 4 Voice 9B rather than being a model trained wholly by Kyutai.
Research release
The training method changes the reasoning channel
The supervised stage stitches written reasoning with speech chunks before reinforcement learning. Not all internal reasoning is spoken. The release still needs the separate speech tokenizer and decoder.glm-4-voice-of-reason-stitch-9b: Kyutai model card
Specifications & access
| Detail | Value and conditions |
|---|---|
| Status | Research release |
| Access | Downloadable |
| Model identifiers | kyutai/glm-4-voice-of-reason-stitch-9b |
| Upstream model | GLM 4 Voice 9B.glm-4-voice-of-reason-stitch-9b: Kyutai model card |
| Terms | The GLM 4 Voice license applies; this is not an Apache-licensed release.glm-4-voice-of-reason-stitch-9b: Kyutai model card |
Related models
- GLM 4 Voice of Reason 9B
Kyutai ยท Voice of Reason
A GLM 4 Voice 9B adaptation trained with reinforcement learning for spoken mathematical reasoning.
Launch date not confirmedResearch releaseDownloadable