KRAFTON AI Open-Sources Raon-1 Mini: A 30B MoE Model Rivaling Meta's Muse Glimmer

1 reports archived1 independent sourcesupdated 10/25/2026, 21:27:18

Synthesis & Latest Updates

1 Sources Cross-Validated

KRAFTON AI has released and open-sourced the preview of Raon-1 mini, a model trained from scratch. Featuring a 30B-A4B Mixture-of-Experts architecture with linear attention and speculative decoding, it is licensed under Apache 2.0. The company claims performance comparable to Meta's Muse Glimmer 30B in coding and agentic tasks, while offering faster inference and better memory efficiency.

LATEST/KRAFTON AI has released and open-sourced the preview of Raon-1 mini, a model trained from scratch. Featuring a 30B-A4B Mixture-of-Experts architecture with linear attention and speculative decoding, it is licensed under Apache 2.0. The company claims performance comparable to Meta's Muse Glimmer 30B in coding and agentic tasks, while offering faster inference and better memory efficiency.

Hourly heat curve

2 fully observed hours

Fewer than 3 fully observed hours; trend not plotted yet.

Coverage timeline

Total 1 reports · Latest first

IT之家 · 智能时代T2Oct 25, 21:27·78 pts
KRAFTON AI Open-Sources Raon-1 Mini: A 30B MoE Model Rivaling Meta's Muse Glimmer
- Raon-1 mini utilizes a 30B-A4B MoE architecture with linear attention and speculative decoding support
- Fully open-sourced under the Apache 2.0 license, with a limited-time free API trial available
- Officially claimed to match Meta Muse Glimmer 30B in coding and agentic tasks, but with faster speed and lower memory usage
View Related Reports & Analysis Original Source