Kokoro-82M

An 82M TTS model that punches absurdly above its weight — ranked #1 on TTS Arena against 7B+ models, runs fine on CPU. The honest default for narration, read-aloud, voice-over.

License: Apache 2.0 · Context: n/a · Released: December 25, 2024

The decision in five lines

The call: Skip for local — for voice
Best for: voice
Runs on: 23 hardware picks fit (cheapest: Intel Arc B580 12 GB · $249)
Watch out: Voice cloning or paralinguistic expression — Kokoro is fixed-voicepack only.
Evidence: Estimated · last verified April 2026

82M (StyleTTS2-based): PARAMETERS
TTS: TYPE
—: CONTEXT
<500 MB (CPU-capable): VRAM AT Q4

Where we recommend this

Every tier slot in the planner where this model is a top or alternate pick. Pulled live from planner.js — when the planner refreshes, this table stays current.

VOICE · LOW

Kokoro-82M (Apache 2.0)Community daily driver for English TTS; CPU-real-time at 82M params; v1.0 with 8 languages and 54 voices. No voice cloning.

The call

An 82M TTS model that punches absurdly above its weight — ranked #1 on TTS Arena against 7B+ models, runs fine on CPU. The honest default for narration, read-aloud, voice-over.
When not to use: Voice cloning or paralinguistic expression — Kokoro is fixed-voicepack only. For cloning use Chatterbox or VoxCPM2.

Runner notes

`pip install kokoro`. Also available as ONNX (`onnx-community/Kokoro-82M-v1.0-ONNX`) for JS/edge. Real-time on CPU with no GPU needed.

License: Apache 2.0
Released: December 25, 2024
Maker: hexgrad
Model card: huggingface.co/hexgrad/Kokoro-82M →

Hardware that fits

Every hardware pick whose memory fits this model at the quant we recommend. Sorted cheapest-first — the top row is your best-value fit. Click through for the full buyer’s guide.

Next step

Find-by-model — see what hardware runs this→