@magnitudedev/cli@0.2.4 @magnitudedev/cli@0.2.5
shadcn-ui/uiOct 2, 2026 → Oct 3, 20261 day between releases
What's New
Patch release adds significant speed boosts for multi-token prediction and prompt processing.
Improvements
- 8-11% speed boost for 16K context on Macs and NVIDIA GPUs.
- 2x speedup for prompt processing on M5 and later Macs via GPU tensor ops.
- Speedup for MoE models by drafting three tokens ahead.
Bug Fixes
- Fixes for M1/M2 Mac kernel loading failures.
- Vulkan shader compilation fixes.
Upgrade Notes
No configuration changes are required for this patch release.
The Two Releases
@magnitudedev/cli@0.2.4
Oct 2, 2026· 20 days ago
This patch improves resource efficiency by reducing memory usage and optimizing kernel tuning times. It also fixes critical bugs related to model loading on older Mac hardware and improves error handling for partial request failures.
Full release notes@magnitudedev/cli@0.2.5
Oct 3, 2026· 19 days ago
This release focuses heavily on performance improvements, including significant speed boosts for multi-token prediction and context handling on various hardware. It also addresses critical stability issues on Apple Silicon (M1/M2) and Vulkan GPUs.
Full release notes