llama.cpp Release b9653 Adds Vulkan Support for More CONCAT Operations and Multi-Platform Binaries
English summary
The b9653 release of llama.cpp extends the Vulkan backend to handle additional CONCAT tensor operation types, improving compatibility for models that rely on these operations. It also ships pre-built binaries for macOS (Apple Silicon, Intel), Linux (multiple GPU backends including Vulkan, ROCm, OpenVINO, SYCL), Android, Windows (CUDA 12/13, Vulkan, SYCL, HIP), and openEuler platforms. The release was published automatically on June 15, 2026.
Chinese summary
llama.cpp 的 b9653 版本扩展了 Vulkan 后端，使其支持更多 CONCAT 张量操作类型，提升了依赖此类操作的模型的兼容性。该版本同时提供了针对 macOS（Apple Silicon、Intel）、Linux（含 Vulkan、ROCm、OpenVINO、SYCL 等多种 GPU 后端）、Android、Windows（CUDA 12/13、Vulkan、SYCL、HIP）以及 openEuler 平台的预编译二进制文件，于 2026 年 6 月 15 日自动发布。
Key points
Vulkan backend now supports more CONCAT operation types, which can unblock previously incompatible model architectures.
Vulkan 后端现在支持更多 CONCAT 操作类型，可解锁先前不兼容的模型架构。
Pre-built binaries are available for a wide range of platforms and GPU backends, including CUDA, Vulkan, ROCm, OpenVINO, and SYCL.
预编译二进制文件覆盖了广泛的平台和 GPU 后端，包括 CUDA、Vulkan、ROCm、OpenVINO 和 SYCL。
The release is an automated build (b9653) triggered by CI/CD workflows, not a manually curated version.
该版本是通过 CI/CD 工作流自动构建的（b9653），而非人工策划的版本。