llama.cpp Releases
684 releases of ggml-org/llama.cpp
- v2.1.6
Minor release featuring improvements to task list labels, development enhancements, and fixes to Docker production setup.
Feb 7, 2026
- v3.4.0
Released PaddleOCR-VL-1.5 complex document parsing solution with SOTA accuracy, added MLX-VLM inference support, and enhanced table/heading reconstruction capabilities.
Jan 29, 2026
- 0.2.0unlimited audio length support
Updated the inference pipeline and model cards to support unlimited audio length processing.
Dec 22, 2025
- 0.1.0
Dec 12, 2025
- v1.1.0
The initial major feature release adding Apple Silicon (MPS) support, multi-GPU support, IBM Granite MoE support, and improvements for Colab/Kaggle notebook compatibility.
Dec 10, 2025
- v1.0.1
Nov 16, 2025
- v2.1.5
This update resolves Docker initialization issues for PostgreSQL and MinIO, adds Korean language support, and updates project dependencies and documentation.
Sep 25, 2025
- v2025.924.154138
A maintenance release focused on build stability, fixing Windows service installation syntax and removing XCode dependency on macOS to streamline the compilation process.
Sep 24, 2025
- v2025.923.33222
A critical security update release that includes fixes for power management, display device options, API endpoints, video settings, and updates dependencies to FFmpeg 8.0 and GCC13+.
Sep 23, 2025
- v2.0.25
This release focuses on fixing various bugs related to configuration storage, packet interception, and UI state preservation, while adding a new zoom shortcut feature.
Sep 5, 2025
- v2.1.4
A feature-rich release focusing on project management enhancements, including user activity feeds, Kanban board improvements, database performance indexing, and various UI/UX refinements.
Aug 14, 2025
- v1.3.1
This minor update introduces EdgeTTS support for dubbing, adds a dark mode to the desktop and server interfaces, includes a configuration page for the server, and optimizes FasterWhisper GPU acceleration.
Jul 27, 2025
- v2.1.3
A major update focused on user experience, featuring a complete overhaul of the authentication system, dark mode support for task management, and significant architectural improvements for code maintainability and performance.
Jul 23, 2025
- v1.5.0SANA-1.5 && SANA-Sprint
Merges SANA-1.5 and SANA-Sprint into a unified codebase, introducing support for high-resolution images (2K and 4K), advanced training techniques like FSDP and LoRA, and enhanced inference capabilities with reduced VRAM requirements.
Mar 25, 2025
- v1.22.22
Minimal release notes: "- Fixes a punycode warning. - Fixes a hoisting issue when transitive dependencies themselves listed aliases as depend"
Mar 9, 2024
- v1.22.21
Fixed an issue where calling Yarn from a project subfolder outside of a Corepack context caused failures.
Nov 14, 2023
- v1.22.20
Embedded Punycode into the bundle and improved the `yarn set version` command to install exact versions.
Nov 14, 2023
- ext2_image
May 16, 2023
- v1.22.19
Minimal release notes: "- Adds compatibility with WebAuthn on the npm registry"
May 10, 2022
- v1.4.4Firefox Snap support for Ubuntu 22.04
Adds support for Firefox Snap on Ubuntu 22.04 and fixes crashes related to Certificate Signing Requests (CSRs).
Apr 26, 2022