How to run CSM-1B (Sesame) locally

Sesame · Text → speech · 1B params · Apache-2.0 (commercial OK)

A conversational speech model from Sesame using a Llama backbone to generate context-aware, natural-sounding dialogue audio.

What it costs to run — $0 markup

On your own machine — $0. Runs free if you have about 8.0 GB of VRAM (~6-8GB; Llama backbone + Mimi codec decoder), via Transformers.
Rent a GPU — from $0.26/hr. Fits RTX 3090 24GB at the direct vendor price ($0 markup) — pay only for the minutes you generate.
Download the weights — free. Open weights at sesame/csm-1b.

Note: generative-media models are billed per image / per second / per minute on hosted services — not per token. Running locally or on your own rented GPU is usually far cheaper and keeps your data on your machine.

Key facts

Does	Text → speech
VRAM to run	~8.0 GB (~6-8GB; Llama backbone + Mimi codec decoder)
Download	~4 GB
Parameters	1B
License	Apache-2.0 (commercial use OK)
Run with	Transformers

Get CSM-1B (Sesame) on Hugging Face →

More voice & audio models

Whisper large-v3 — Speech → text, ~10 GB VRAM
Whisper large-v3-turbo — Speech → text, ~6.0 GB VRAM
Kokoro-82M — Text → speech, ~1.0 GB VRAM
XTTS-v2 — Voice cloning, ~6.0 GB VRAM
MusicGen Large — Text → music, ~16 GB VRAM
Chatterbox — Text → speech, ~8.0 GB VRAM

Browse: all media models · chat / LLM models