Kokoro Models by Hexgrad

HexgradApache 2.0Open sourceAudio
1 model2025From $0.620/1M input

Last refreshed 2026-11-18. Next refresh: weekly.

Details

ResearcherHexgrad
LicenseApache 2.0OSI-approved
Commercial usePermitted
Models1
Released2025

Capabilities

MultimodalAll models

About

Kokoro is hexgrad’s family of open-weight text-to-speech models, published on Hugging Face.

Decision facts

Best fit
audiotext to speechvision and multimodal work
Capability starting point
Kokoro-82M with multimodal inputs
Lowest tracked input
Kokoro-82M · $0.620/1M · OpenRouter
Closest related family
AssemblyAI

Current Variants

Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.

1 in view
Kokoro-82MCurrent

Use when the workload needs text to speech, 82M parameters, and multimodal inputs.

2025-01text to speech82M parametersmultimodal inputs

Release Timeline

1 release group
2025-01
1 current
Kokoro-82M
text to speech82M parametersmultimodal inputs
Current

Specifications(1 models)

Kokoro model specifications comparison
ModelReleasedParametersMultimodal
Kokoro-82M2025-0182MYes

Available From(4 providers)

Pricing

Kokoro model pricing by provider
ModelProviderInput / 1MOutput / 1MType
Kokoro-82MOpenRouter$0.62—Serverless