{"id":3056,"date":"2026-09-24T06:13:35","date_gmt":"2026-09-23T21:13:35","guid":{"rendered":"https:\/\/localmodelwatch.tsuchitsuchi.com\/2026\/09\/24\/llama-cpp-v0-5-0-released\/"},"modified":"2026-09-24T06:13:35","modified_gmt":"2026-09-23T21:13:35","slug":"llama-cpp-v0-5-0-released","status":"publish","type":"post","link":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/09\/24\/llama-cpp-v0-5-0-released\/","title":{"rendered":"llama.cpp v0.5.0 Released with Backend and Server Upgrades"},"content":{"rendered":"

<\/p>\n

At a Glance<\/h2>\n
\n\n\n\n\n\n\n\n\n\n
Item<\/th>\nValue<\/th>\n<\/tr>\n<\/thead>\n
Repository<\/td>\nggml-org\/llama.cpp<\/a><\/td>\n<\/tr>\n
Version<\/td>\nv0.5.0<\/a><\/td>\n<\/tr>\n
Published<\/td>\n2026-09-24<\/td>\n<\/tr>\n
License<\/td>\nMIT<\/td>\n<\/tr>\n
Source type<\/td>\nPrimary source (the publisher itself)<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n

Values determined by this site’s code at collection time. Dates are JST.<\/em><\/p>\n

<\/p>\n

Overview<\/h2>\n

v0.5.0, the latest version of the C\/C++ LLM inference library llama.cpp<\/code>, has been released. This update focuses on improving backend performance, expanding model support, and enhancing the robustness of server and router operations.<\/p>\n

The most impactful changes for users are hardware acceleration enhancements accompanying backend (ggml) updates, such as accelerated conv2d<\/code> via implicit GEMM in CUDA and fused optimization for MoE and SSM_CONV in Metal. These are expected to improve inference performance under specific environments.<\/p>\n

Key Changes<\/h2>\n

Backend and Hardware Acceleration Optimization<\/h3>\n

In CUDA environments, conv2d<\/code> operations have been accelerated using implicit GEMM. Additionally, in Metal environments, fusion optimization for MoE (Mixture of Experts) and SSM_CONV has been added. These changes directly affect inference speeds for users utilizing NVIDIA GPUs or Apple Silicon.<\/p>\n

Server Feature Expansion and Multi-Address Support<\/h3>\n

In llama-server<\/code>, the --host<\/code> option can now accept comma-separated TCP addresses and UNIX sockets. This enables binding to multiple network interfaces simultaneously\u2014for example, listening to connections via VPN and localhost at the same time. This is a very useful feature for engineers operating servers across multiple network environments.<\/p>\n

Support for New Model Architectures<\/h3>\n

Support for numerous new models and architectures has progressed. Specifically, this includes support for HRM-Text (DFM Mimir 1B), conversion support for MiMo-V2.6, and DFlash (speculative decoding) support for HunyuanOCR. Furthermore, expanded support for Nemotron MTP and implementation of hyper-connection operations and sparse flash attention for Qwen4Exp have also been carried out. Users wanting to try new models locally can now use them in GGUF format thanks to these additions.<\/p>\n

Multimodal Support in Server Output<\/h3>\n

Support for input_image<\/code> has been added to server function-calling outputs. This enables the construction of more advanced multimodal agent functions, such as handling image data as the result of tool execution. This affects developers utilizing multimodal models like Gemma 4 to process complex tasks involving images.<\/p>\n

Sampling Control via Environment Variables<\/h3>\n

Settings for temperature<\/code>, top-p<\/code>, min-p<\/code>, and various penalties can now be configured via environment variables. This allows users to control sampling parameters simply by setting execution environment configurations without rewriting code. Convenience is improved for users who want to dynamically adjust inference behavior.<\/p>\n

Supported Models and Hardware<\/h2>\n

This release enhances support for the following model architectures and formats:<\/p>\n