ggml-org/llama.cpp is a C/C++-based LLM inference project, and release b11988 added support for Limite-series 1B models. This change was reflected on both the conversion script and the runtime side. This is the change stated in the repository description and release summary for this build. The scope of this news item is therefore limited to support for the new model series, and there is nothing in the FACTS saying that other features or improvements were included.
Mentioning the conversion script and the runtime together means that this support is not limited to model conversion but also covers the execution side. It is not that only one part of the model support was added on its own; the release summary names both areas as targets of the change. This is the concrete change confirmed in b11988, and it targets Limite-series 1B models. However, the detailed scope of the support cannot be narrowed any further from this summary alone.
The FACTS do not give specific model names in the Limite series, model file formats, execution conditions, or performance measurement results. Therefore, which files or settings to use the support with, and how speed or results differ compared with existing models, cannot be confirmed from this material. The source of the change is the b11988 release of ggml-org/llama.cpp. The only conclusion available for now is limited to the scope that support was added to both the conversion script and the runtime.