llama.cpp Releases
609 releases of ggml-org/llama.cpp
- b10167
Refactored memory management calls by abstracting `llama_memory` to a common `common_memory` function to improve code structure and potential reuse.
Jul 28, 2026
- b10166
This release focuses on internal graph optimizations in the ggml library, specifically addressing how view sources and outputs are handled to improve graph correctness and sampler stability.
Jul 28, 2026
- b10165
Restores support for IQ4_NL quantization format in the Vulkan backend's Fully Attention (FA) mode to fix compatibility issues.
Jul 28, 2026
- b10164
Accelerates Mamba-2 prefilling operations using a chunked SSD (Sliding Window) matrix multiplication implementation in CUDA.
Jul 28, 2026
- b10159
Implements a Fast Walsh-Hadamard Transform (FWHT) kernel for the Metal backend on Apple Silicon to optimize performance.
Jul 28, 2026
- b10158
Adds support for the Eagle3-V3 specification used in GPT-OSS models, enabling better compatibility with these architectures.
Jul 28, 2026
- b10156
Disables the -ffast-math compiler flag for HIP (Radeon) support to improve numerical precision in calculations.
Jul 28, 2026
- b10155
Adds support for the MiMo-V2.5 audio model, which utilizes Residual Vector Quantization (RVQ) for audio input processing.
Jul 27, 2026
- b10154
Introduces a new utility function to print available computing devices, aiding in device detection and configuration.
Jul 27, 2026
- b10153
Adds support for the Nanbeige4.2 model architecture, expanding the range of supported models in the library.
Jul 27, 2026
- b10152
Optimizes GPU layer allocation for Multi-Threaded Parallel (MTP) blocks to ensure the front layers of the model remain on the GPU when using the n_gpu_layers parameter.
Jul 27, 2026
- b10151
Parallelized the OpenCL compiler (ocloc) invocations in the SYCL build process for better build performance.
Jul 27, 2026
- b10150
Adjusted the logic for offloading operations to the backend of the weights, and fixed graph issues for DeepSeek-V4.
Jul 27, 2026
- b10149
Removed an unnecessary synchronization call in the state save/load test to improve performance.
Jul 27, 2026
- b10148
Fixed a regression where explicitly specified draft models were silently overridden by automatic sidecar resolution, ensuring user configuration takes precedence.
Jul 27, 2026
- b10146
Enabled BF16 (Brain Float 16) tiled GEMM optimizations specifically for PowerPC architecture.
Jul 27, 2026
- b10145
Modified the --mlock flag behavior to exclusively lock memory in place without memory mapping.
Jul 27, 2026
- b10144
Improved the web server and UI stability by fixing stream handling for model names with special characters, managing pending requests during model loading, and preserving session state on page reloads.
Jul 27, 2026
- b10142
Added preliminary support for the MiniMax-M3 multimodal model, including vision capabilities, optimized attention mechanisms, and multi-stream handling.
Jul 27, 2026
- b10141
Fixed a build issue for the multimodal text module (mtmd) on Android.
Jul 26, 2026