Model Terminal

Kokoro-82M

Kokoro-82M is an open-weight, Apache-licensed text-to-speech model with 82 million parameters, designed for fast and cost-efficient speech synthesis across multiple voices and languages. It is published on Hugging Face by the independent researcher account hexgrad. Source: supabase.

Value score
Not scored
Context
tokens
Max output
tokens
Price

Capability radar

Peer value bars

Identity

Developer
Hugging Face
Openness
Modalities
Text-to-speech
Release
Knowledge cutoff
Deprecation
API docs

Benchmarks

No benchmark scores yet.

Compare nearby

Pricing

Input / 1M
Output / 1M
Speed