mlxrunner

This release introduces Qwen3.8 Flash Next support for MLX and adds structured output capabilities to mlxrunner. It also includes improvements to cmake compatibility patches and addresses Metal GPU timeouts.

This release brings updates to MLX and llama.cpp, adds structured output support to mlxrunner, and improves GPU timeout handling.

This release introduces the Claude desktop app and enhances the user experience with polished onboarding and a new 'Connect your apps' feature. It also includes several bug fixes for MLX and the DeepSeek Harness.

This release introduces several new features for the desktop application, including Claude support and a "Connect your apps" experience. It also includes bug fixes for the MLX runner and general linting improvements.

This release introduces new auto-installation features for models like Claude Code and opencode, alongside numerous stability and performance improvements across GPU handling (Vulkan, CUDA presets) and model loading/generation.

This release introduces new auto-installation features for models like Claude Code and opencode, alongside numerous performance and stability fixes across Vulkan, speculative decoding, and model offloading mechanisms. It also updates the underlying llama.cpp dependency.

This release includes improvements to MLX MTP caching, hardening of mlxrunner layers, and a fix for launch provider drift. Prompt caching logic has also been decoupled from context shifting.

This release introduces model batching support and adds NVIDIA TensorRT Model Optimizer import capability. Several minor bugs related to tokenization and desktop app session handling were also resolved.

This release introduces model batching support and TensorRT Model Optimizer import for the mlx backend. It also includes several bug fixes related to tokenization and desktop application startup behavior.

This release introduces model batching support and fixes several issues related to tokenization and desktop application startup behavior. It also includes support for NVIDIA TensorRT Model Optimizer import.

This release introduces support for NVIDIA TensorRT Model Optimizer import within mlx and fixes an issue related to multi-regex BPE offset handling in the tokenizer. It also includes performance improvements by batching the sampler across multiple sequences in mlxrunner.

This release focuses on stability improvements, including fixing a memory leak in mlx and adjusting settings for the Grok model on ggml. It also updates VS Code documentation and hides the VS Code launch option.

This release introduces debug request logging and improves MLX performance with better cache sharing and new format imports. Several stability fixes were also implemented across the desktop app, MLX runner, and CI.

This release introduces debug request logging, improves KV cache sharing in mlxrunner, and fixes several stability issues including desktop app loading hangs and mlxrunner deadlocks.

This release focuses on stability and performance improvements across parsers, cloud proxy handling, and MLX backend optimizations. It also includes fixes for Docker builds and application defaults.

This release introduces support for the Nemotron architecture and includes several performance and stability improvements, particularly around MLX memory usage and logging. It also updates the mlx-c bindings.

This release introduces support for the nemotron architecture and includes several performance and logging improvements, particularly for MLX-based operations. It also updates underlying MLX-C bindings.

This release introduces support for the nemotron architecture and includes several performance and logging improvements, particularly for MLX-based operations.

This release introduces support for several new model architectures (Gemma 3, Llama 3, Qwen 3) in mlxrunner and adds the new `ollama launch` CLI command. Several minor bug fixes related to mlx model display and scheduling were also implemented.