LIVE
Formalizing Fermat's Last Theorem04/09/26 · Anthropic|OpenAI and Anthropic suffer simultaneous outages with no official explanation04/09/26|Show HN: Open-Source eInk Bike Computer04/09/26|Corporate America increasingly turns to open-source AI04/09/26|Study finds Google's AI Mode surfaces pricier products than classic search04/09/26 · Google|Study measures how coding agents pick third-party tools03/09/26|OpenAI's GPT-6 Astra sets new state-of-the-art on ARC-AGI-303/09/26 · OpenAI|ESPO: A Prompt Optimization Method That Beats GEPA With Shorter, More Stable Prompts03/09/26|"Last Translation Benchmark": a large-scale collaborative effort for a definitive MT benchmark03/09/26|NVIDIA pushes local AI at IFA 2026 with RTX Spark PCs and a personal inference router03/09/26 · NVIDIA|Google DeepMind unveils WeatherNext 3, its most accurate weather AI model yet03/09/26 · Google DeepMind|OpenAI Launches $1B Daybreak Program for Frontline Cybersecurity Defenders03/09/26 · OpenAI|Formalizing Fermat's Last Theorem04/09/26 · Anthropic|OpenAI and Anthropic suffer simultaneous outages with no official explanation04/09/26|Show HN: Open-Source eInk Bike Computer04/09/26|Corporate America increasingly turns to open-source AI04/09/26|Study finds Google's AI Mode surfaces pricier products than classic search04/09/26 · Google|Study measures how coding agents pick third-party tools03/09/26|OpenAI's GPT-6 Astra sets new state-of-the-art on ARC-AGI-303/09/26 · OpenAI|ESPO: A Prompt Optimization Method That Beats GEPA With Shorter, More Stable Prompts03/09/26|"Last Translation Benchmark": a large-scale collaborative effort for a definitive MT benchmark03/09/26|NVIDIA pushes local AI at IFA 2026 with RTX Spark PCs and a personal inference router03/09/26 · NVIDIA|Google DeepMind unveils WeatherNext 3, its most accurate weather AI model yet03/09/26 · Google DeepMind|OpenAI Launches $1B Daybreak Program for Frontline Cybersecurity Defenders03/09/26 · OpenAI|
🇺🇸 United States🎙️ Voice synthesisReleased December 2024

Kokoro-82M

Hexgrad
Open-sourceSelf-hostRGPD
Verdict

A solid pick for self-hosting and full data control.

Overview

Kokoro-82M is an open-weight text-to-speech model with just 82 million parameters, built on a decoder-only StyleTTS2/ISTFTNet architecture with no diffusion step. It covers 8 languages and 54 voices, trained on only a few hundred hours of permissively licensed audio, which caps its expressiveness and language depth compared with proprietary TTS systems trained on far larger corpora. Released under Apache 2.0, it runs comfortably on a laptop, and once self-hosted keeps all voice data on the user's own infrastructure regardless of the vendor's US origin.

Skill profile

Not disclosed

Strengths

  • Open-source and self-hostable

Limitations

  • API pricing not disclosed

Who is it for

A good fit if…
  • you want to control cost or self-host
  • you have GDPR constraints
Skip it if…

    Ideal use cases

    • Rapid voice prototyping
    • Embedded voice assistants
    • Automated voice-over generation
    • Text-to-speech accessibility
    • Local, cloud-free deployments

    Access & availability

    Self-hostable (open weights)

    Key specifications

    Context
    0
    Input price
    Output price
    Speed
    Price not auditedBenchmarks not audited

    Privacy

    GDPR-compliantHosting : self

    Run it locally

    Runs on a laptop11.6M7k
    QuantizationDiskRAM / VRAMTypical hardware
    Q4 · recommended0 GB2 GBAny recent PC/Mac
    Q8 · balanced0.1 GB2 GBAny recent PC/Mac
    FP16 · max quality0.2 GB2 GBAny recent PC/Mac

    Estimates for a moderate context. Long contexts need more RAM (KV cache).

    Deploy

    Copy-ready commands generated from this card. Adjust context length and GPU count to your hardware.

    Estimated memory · FP16 2 Go · Q4 2 Go

    OpenAI-compatible server for production on NVIDIA GPUs.

    pip install vllm
    vllm serve hexgrad/Kokoro-82M \
      --max-model-len 32768 \
      --tensor-parallel-size 1 \
      --dtype auto
    hexgrad/Kokoro-82M
    Advanced data · for experts
    Architecture, modalities, detailed cost, full benchmarks
    Architecture
    dense
    Size
    82M
    Cutoff date
    Inputs
    text
    Outputs
    text
    License
    apache-2.0
    Hosting
    self
    Hallucination score
    Estimated cost (API)
    Typical exchange (~3k in / 1k out)
    Not disclosed
    1M in + 1M out
    Not disclosed
    All benchmarks
    Arena Elo
    MMLU
    GPQA
    HumanEval
    SWE-Bench
    MATH

    Similar models