model-support

Support for the new Ling 3.0 reasoning models has been integrated into llama.cpp, making them accessible via this development tool.

The article details the current support status for the Ling-3.0-flash model weights across inference engines: SGLang commits to day-0 integration, vLLM awaits open weights, and llama.ccp lacks conversion for the Bailing MoE variant. It notes that the release pattern involves a free API window followed by open-sourcing, as seen with Ling-2.6-flash.

This pull request adds support for the Hy3 (hy_v3) model with MTP speculative decoding to llama.cpp, enabling efficient inference for this architecture.

GPT-5.6 is now supported in Hermes Agent and available via the Nous Portal, providing access to the new model.

Pull request adding support for DeepSeek V4 model in llama.cpp, enabling inference of this model on various hardware.

llama.cpp adds support for granite-speech-4.1-2b and LFM embedding models, introduces multiple Vulkan backend enhancements (e.g., CONV_3D, spec constants, overflow fixes), and includes UI/UX improvements and miscellaneous fixes.

Claude Fable 5 is now supported in Hermes Agent via Nous Portal, with a promotion offering the first 500 new users one month free access to the Plus plan.

Pull request adding support for StepFun 3.5 MTP model in llama.cpp.

A pull request has been merged into llama.cpp to add support for the Mimo v2.5 model, enhancing the framework's compatibility with this specific AI architecture.