llama.cpp has released build b11338, focusing on stability and performance improvements for Qualcomm Hexagon NPUs. The update implements shared strided DMA copies for CPY and CONCAT operations, removing the previous broken logic that restricted DMA usage based on row dimensions. Additional fixes include adding missing dma_queue_flush() calls and implementing guards for unsupported conditions to prevent runtime errors. The release maintains support for CUDA 12/13, ROCm 10.0, Vulkan, and various CPU architectures across macOS, Linux, Windows, and Android.
Read originalEarlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
llama.cpp Releases · September 20, 2026 · Same story
llama.cpp Releases · September 30, 2026 · Same story
This release quietly expands llama.cpp's hardware reach with two major additions: ROCm 10.0 for AMD GPUs and native support for Linux arm64 Snapdragon devices. The inclusion of ROCm 10 is significant, as it brings AMD users closer to parity with CUDA in terms of supported versions, reducing the friction for local inference on non-NVIDIA hardware. Meanwhile, Snapdragon support opens up a new class of mobile AI acceleration, allowing developers to leverage Adreno GPUs and Hexagon NPUs directly. While Apple Silicon builds have KleidiAI disabled by default, the core value here is the broadening of accessible compute backends without requiring complex custom compilation.
This release quietly cements llama.cpp as the universal inference runtime by finally bringing ROCm 10.0 to Linux and Windows alongside CUDA 13.4, effectively closing the hardware gap for AMD users who previously lagged behind NVIDIA. The standout addition is native support for Linux arm64 Snapdragon devices, enabling local AI on mobile-class silicon with CPU, Adreno GPU, and Hexagon NPU acceleration. While KleidiAI on Apple Silicon is currently disabled in this build, the broader expansion to diverse accelerators means developers no longer need to compile from source to target non-NVIDIA hardware. The world now has a single binary ecosystem that runs everywhere from x86 servers to ARM mobile chips.
This release addresses a memory management issue in the Metal backend by properly releasing temporary private transfer buffers. While not a feature upgrade, it stabilizes performance on Apple Silicon devices where previous builds might have suffered from memory pressure or fragmentation during inference. The update also includes standard platform binaries for CUDA 13 and ROCm 10.0, ensuring compatibility with the latest NVIDIA and AMD driver stacks. For local LLM users on macOS, this is a quiet but necessary maintenance step to keep inference smooth.
Anthropic quietly fixed a critical credential leakage bug where MCP error messages were exposing raw API keys in plaintext logs. Beyond the security patch, this release stabilizes the notoriously fragile background agent system by fixing subagent hand-offs and connection stalls that previously caused silent failures. The update also tightens session management for cloud environments, ensuring large transcripts actually load instead of hanging indefinitely. It’s a maintenance-heavy release, but essential for anyone running complex, multi-step automated workflows.
This update shifts Claude Code from a simple CLI wrapper to a more extensible platform by introducing 'Claude Mods,' allowing plugins to modify deeper behavior rather than just adding tools. The inclusion of a built-in 'You should know' side agent that flags potential oversights is a notable step toward autonomous oversight within the coding workflow. Beyond features, the release addresses critical stability issues in remote sessions and significantly improves accessibility for screen reader users, making the tool more robust for enterprise and diverse developer environments.
The v0.31.0rc3 release of vLLM brings a critical infrastructure tweak to the new Model Runner V2: support for randomized dummy inputs. This isn't a feature for end-users but a developer-facing fix that stabilizes how the runner handles initial tensor shapes during compilation and warm-up phases. By allowing randomized inputs, it reduces the likelihood of shape-mismatch errors when tracing models with dynamic dimensions. For builders running large-scale inference workloads, this means fewer silent failures and more robust model loading sequences in production environments.