The Open Weights
LatestModelsLeaderboardsCompanies
Subscribe
The Open Weights

The daily record of open-source AI. New model releases, leaderboards, and what's coming next — written for people who ship.

Refreshed every 12 hours

Discover

  • Latest releases
  • New today
  • Trending models

Browse

  • All models
  • Companies
  • Categories
  • Leaderboards

About

  • About
  • Editorial policy
  • RSS feed
  • Newsletter

© 2026 The Open Weights. An independent publication.

PrivacyTermsSMSAggregated by Claude · curated by humans.
LatestKRAFTON1.0
KRAFTONAny-to-Any

KRAFTON Releases 9B Bilingual Speech Model

The gaming giant behind 'PUBG' has released Raon-Speech-9B, a multimodal model for English and Korean speech recognition and synthesis.

Mar 30, 2026
NotableCC BY-NC 4.0
Raon-Speech-9B

South Korean game developer KRAFTON, best known for the hit title PUBG: Battlegrounds, has released a new open-source AI model focused on audio. The new model, named Raon-Speech-9B, is a 9-billion-parameter system designed for a range of multimodal speech tasks.

The model is bilingual, supporting both English and Korean. Its core capabilities include automatic speech recognition (ASR), which converts spoken language to text, and text-to-speech (TTS) synthesis for generating spoken audio from text. This dual functionality makes it a versatile tool for building sophisticated voice-based applications.

Raon-Speech-9B is a notable release from a company primarily known for gaming, highlighting a broader trend of specialized industries contributing significant models to the open-source community. Its strong support for Korean is also a welcome addition, providing high-quality resources for a language often underserved by major model releases.

KRAFTON has made the model available for research and experimentation on its Hugging Face repository. It is released under the Creative Commons Attribution-NonCommercial 4.0 license, which allows for broad non-commercial use and adaptation.

Sources

  • KRAFTON/Raon-Speech-9B

    Hugging Face

    Visit

Get the model

Hugging Face

Specs

Parameters9B
Size18.1 GB
PrecisionBF16
ArchitectureRaonModel
LicenseCC-BY-NC-4.0
Downloads3.4K
Likes57

Modalities

Any-to-AnyText → SpeechSpeech → Text

0 comments

No comments yet. Be the first to weigh in.

More in Any-to-Any

StepFun/Text → Speech

StepFun's StepAudio 3 Realtime targets live voice AI

The audio-language foundation model builds a listen-converse-think-act loop aimed at natural, low-latency spoken interaction.

Sep 11, 2026
SenseTime/Any-to-Any

SenseTime's SenseNova-U1.5 Unifies Vision Tasks

An 8B model drops the usual encoder and VAE in favor of a single native architecture spanning understanding, reasoning, and image generation.

Sep 9, 2026
SenseNova-U1.5-8B-MoT
SenseTime/Any-to-Any

SenseTime Releases SenseNova U1.5 8B Any-to-Any Model

The new 8B multimodal model handles text, images, and image editing within a single native architecture.

Aug 19, 2026