Coding Tools
llama.cpp b11183 splits Metal FA kernels
llama.cpp Releaseshigh confidence
Why it matters
- →Refactoring Metal FA kernels optimizes memory usage for Apple Silicon users.
- →Adds ROCm 10.0 and CUDA 13.4 support, keeping AMD/NVIDIA parity current.
- →KleidiAI is disabled on Apple Silicon, signaling ongoing stability work.
llama.cpp has released build b11183, featuring a significant internal refactor of Flash Attention kernels for Apple's Metal backend. The change splits these kernels into per-dtype libraries to optimize memory usage and compilation efficiency. The release also updates backend support, adding ROCm 10.0 and CUDA 13.4 builds across Linux and Windows platforms. Notably, KleidiAI support on Apple Silicon is currently disabled in this build. This update focuses on backend stability and hardware compatibility rather than new model architectures or inference speed improvements.
Read originalMore from llama.cpp Releases
Coding Toolscoding
llama.cpp b11184 expands Metal support to wider blocks
This release tackles a specific performance bottleneck on Apple Silicon by extending Metal FWHT kernels to handle block widths up to 8192. Previously limited to 512, these wider operations now use threadgroup memory instead of registers, enabling faster inference for larger context windows or model architectures that rely on Hadamard transforms. The update also ships binaries for CUDA 13 and ROCm 10.0, keeping the toolkit aligned with the latest NVIDIA and AMD driver ecosystems. It’s a quiet but necessary optimization that improves throughput on M-series chips without changing the user experience.
llama.cpp Releases
Coding Toolscoding
llama.cpp b11185 adds ROCm 10 and CUDA 13
One of llama.cpp's quietest but biggest releases. With ROCm 10.0 added as a default build, AMD GPU users stop being second-class citizens for local inference — the gap with CUDA narrows visibly. Apple Silicon Macs now compile in KleidiAI by default, meaning every M-series machine gets ARM-tuned GEMM kernels for free, no flag-flipping required. There's no new model and no new quantization here — just llama.cpp quietly becoming the inference runtime for everyone who isn't on NVIDIA. That's the headline.
llama.cpp Releases
Coding Toolsother
llama.cpp fixes LFM2 audio transcription accuracy
A subtle but critical fix in llama.cpp’s mtmd module corrects the mel preprocessor for the LFM2 audio model, eliminating transcription errors that affected over half of tested Japanese utterances. By switching to a symmetric Hann window and adjusting normalization epsilon, the build now matches reference implementations with near-zero error rates instead of changing entire words. This resolves a significant reliability gap for local multilingual speech-to-text, ensuring that open-source inference produces results comparable to proprietary baselines without manual intervention.
llama.cpp Releases
More in Coding Tools
Coding Toolscoding
Claude Code v2.1.282 patches critical session and UX bugs
This release stabilizes Claude Code by fixing a cascade of session-breaking errors that previously caused silent data loss or API drops. The most significant fix addresses resumed conversations re-sending messages in altered forms, which was corrupting reasoning traces and breaking extended thinking workflows. It also resolves persistent login refresh loops and managed setting parsing failures that plagued enterprise deployments. While the changelog is dense with UI tweaks like scrollbar fixes and vim mode corrections, the core value lies in restoring reliability for long-running agent sessions.
Claude Code Releases
Coding Toolscoding
Claude Code v2.1.283: Gateway hints and plugin fixes
This release quietly solves a major pain point for enterprise AI workflows by adding gateway hint headers, allowing LLM gateways to correctly group requests per user prompt instead of treating them as isolated events. The new managed settings for availableModelsMatch and deniedModels give organizations precise control over model access, blocking specific versions even when broader allowances exist. Beyond governance, the update stabilizes the plugin ecosystem with rigorous validation checks that prevent silent failures from broken or misconfigured extensions. These changes shift Claude Code from a developer tool to a manageable enterprise component.
Claude Code Releases
© Matt WolfeCoding Toolsagents
OpenAI Introduces GPT-Live and Agent API
OpenAI launches GPT-Live 1 and a new Agent API, enabling real-time voice interactions and autonomous agent development.
Matt Wolfe