Raon-1 nano

Released
2026-11-11
Last refreshed
2026-11-11
Status
Researched 1d ago
Open sourceCommercial use: permittedRAGAgentsLong contextJSON / Tool use

Raon-1 nano is a released rag, agents, and long context model with open-source and 256K context; evaluate it while provider pricing coverage matures.

KRAFTON releases · 2 in the last 12 monthsChangelog →
Specifications
Family
Raon
Released
2026-11-11
Context
256K
Parameters
3.1B (1.5B effective)
Architecture
Decoder Only
Specialization
general
Openness
Open source
License
Apache 2.0OSI-approvedPermitted
Weights
Available
Code
Unknown
Training
Pretrained
Created by
KRAFTON

KRAFTON is a South Korean video game company whose KRAFTON AI unit builds its own models, including the Raon family of o

Website
Pricing

No tracked provider pricing is available yet.

Links
Website HuggingFace

About

Raon-1 nano is a dense on-device language model from KRAFTON AI, released November 11, 2026 under Apache 2.0. It has 1.5B effective parameters (3.1B total, including a 1.5B per-layer embedding table) and natively supports up to 262,144 tokens of context. KRAFTON pruned and distilled Raon-1 mini preview (30B-A4B) with its MoE-to-Dense approach, then trained it further on 3T tokens followed by SFT, OPD and RL. It is designed for thinking mode, natural conversation, instruction following and tool calling in Korean and English.

Providers

No active provider routes are tracked for this model yet.

Capabilities

ReasoningJSON / Tool use

Benchmark peer barsfor RAG

No task-mapped benchmark peers are available for this model yet.

Migration checks

No linked migration route is available for this model yet.

API versions

KRAFTON/Raon-1-nano