llama.cpp adds Nemotron support; KRAFTON AI previews Raon-1 mini model

The llama.cpp project has released version b11819, which includes support for non-standard GQA, BPE tokenizers, and domyn-small pre-types, specifically for the Nemotron model. This update also refines GGUF metadata handling for Metaspace prepend schemes. Separately, KRAFTON AI has released a preview of their Raon-1 mini model, a 30B-parameter Mixture-of-Experts model trained on multiple languages with a focus on Korean. The Raon-1 mini model is available under the Apache 2.0 license and demonstrates strong performance in reasoning, instruction following, and agentic tasks, with support for speculative decoding and quantization. AI

IMPACT Expands tooling for running diverse LLM architectures and provides a new multilingual MoE model for research and development.

RANK_REASON Updates to open-source LLM inference libraries and a preview release of a new MoE model.

Read on llama.cpp — Releases →

• byte-pair encoding
• domyn-small
• GGUF
• GQA
• KRAFTON AI
• llama.cpp
• Nemotron
• Raon-1 mini
• SGLang
• transformers
• vLLM

• model release
• product

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →