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
  • Hardware estimates
  • RSS feed
  • llms.txt
  • Newsletter

© 2026 The Open Weights. An independent publication.

PrivacyTermsSMSAggregated by Claude · curated by humans.
LatestKRAFTON1-mini-preview
KRAFTONText / LLM

KRAFTON releases Raon-1-mini preview under Apache 2.0

The Korean game studio behind PUBG puts out a compact, permissively licensed language model as an early preview.

Oct 21, 2026
UpdateApache 2.0
Raon-1-mini (preview)

KRAFTON, the South Korean game company best known for PUBG, has quietly stepped into open-weight language models with Raon-1-mini-preview, a compact text model released on Hugging Face under the permissive Apache 2.0 license.

As the name suggests, this is a preview rather than a finished product. KRAFTON has positioned Raon-1-mini as a small, text-only model, though the published record leaves key details — parameter count, context window, and benchmark results — unstated for now. That makes this more of a signal of intent than a drop-in replacement for established small models.

Why it matters

Game studios are not the usual source of foundation models, and KRAFTON's entry is a reminder that the open-weight field is broadening beyond the familiar AI labs. A few reasons this is worth watching:

  • The Apache 2.0 license allows commercial use and modification, lowering the barrier for developers to experiment.
  • Compact models are increasingly favored for on-device and latency-sensitive applications, including games.
  • A "preview" label suggests more complete releases, and possibly larger siblings, may follow.

For now, the practical takeaway is modest: developers curious about KRAFTON's approach can download and test the weights today. The more interesting question is whether Raon grows into a sustained model family, and what a company with deep interactive-entertainment experience chooses to optimize for.

Sources

  • KRAFTON/Raon-1-mini-preview

    Hugging Face

    Visit
OlderQwen-Audio-3.1-ASR Adds Speaker-Aware TranscriptionQwen · Alibaba · Speech → Text · 3 days agoNewerRecursive releases Flint, an Apache-2.0 chat modelRecursiveSI · Text / LLM · yesterday

Get the model

Hugging Face

Specs

Context window262K tokens
Size31.9 GB
PrecisionFP8
ArchitectureRaon1MiniForCausalLM
LicenseAPACHE-2.0
Downloads249
Likes132

Can you run it?

Runs on a 24 GB GPU at 4-bit.

  • FP8 (as published)

    48 GB GPU (RTX 6000) or 2×24 GB · 64 GB Mac

    33.1 GB
  • 4-bit

    24 GB GPU (RTX 3090 / 4090) · 32 GB Mac

    18.5 GB
Your machine
GGUF builds
  • KV cache sized for 8,192 tokens of context; longer prompts need proportionally more.

Based on Hugging Face weights metadata. Assumes an 8K context and ~1 GB runtime overhead; actual needs vary. How we estimate


Modalities

Text / LLM

The Weekly Weights

Every open release that mattered, one email a week.

0 comments

No comments yet. Be the first to weigh in.

More from KRAFTON

All KRAFTON releases →
A.X-K2 Raon Speech 21B-A3B
KRAFTON/Any-to-AnyRuns on a laptop

KRAFTON releases A.X-K2 Raon speech MoE model

The game maker's new open model blends text-to-speech and speech recognition in a single 21B mixture-of-experts system with just 3B active parameters.

Jul 27, 2026
Raon-OpenTTS-1B
KRAFTON/Text → SpeechRuns anywhere

KRAFTON releases 1B zero-shot voice-cloning TTS

Raon-OpenTTS-1B is a flow-matching diffusion transformer that clones English voices from short reference clips.

May 21, 2026
Raon-Speech-9B
KRAFTON/Any-to-AnyRuns on a laptop

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

More in Text / LLM

All Text / LLM →
Flint
RecursiveSI/Text / LLMConsumer GPU

Recursive releases Flint, an Apache-2.0 chat model

The new conversational LLM ships under a permissive license with community GGUF builds already available for local use.

Oct 21, 2026
DeepSeek-V4.1-Flash (1017)
DeepSeek/Vision-LanguageDatacenter

DeepSeek V4.1-Flash Adds Vision and Reasoning

The new MIT-licensed checkpoint folds image understanding and reasoning into DeepSeek's fast, mixture-of-experts Flash line.

Oct 17, 2026
Nemotron Labs Dory 1.0 4B
NVIDIA/Text / LLMRuns on a laptop

NVIDIA releases compact Dory 1.0 4B text model

The new Nemotron Labs model packs a 4-billion-parameter language model trained on NVIDIA's Nemotron datasets.

Oct 16, 2026