Overview

CSM-1B is the open base model behind Sesame’s voice research. It uses a Llama backbone and an audio decoder producing Mimi audio codes.

Highlights

  • 1B parameters
  • Context-aware speech generation
  • Multiple speakers
  • Apache 2.0 license

Typical uses

Research and prototyping of conversational voices.

Things to consider

  • Base generation model only, not a full assistant.
  • Only clone voices with consent.

Specifications are taken from the model card on Hugging Face and the provider’s announcement. Benchmark results depend on evaluation settings; see how AI benchmarks work.