# mlx gemv_wide bf16 small-M kernel in mlx 0.32.1 > Already covered: `gemv_wide` (PR 3888, "metal: add gemv_wide for fp16/bf16 matmuls of a few rows") is listed in the MLX 0.32.1 release notes, and its eligibility, routing and measured speedups are recorded in mlx-gemv-wide-bf16-fp16-few-row-matmul-kernel-pr.md. Parent: [Mac local LLMs: Speculative decoding and MTP](https://llms-explorer.com/tree/mac-local-llms-speculative-decoding-and-mtp/) · 1 facets · 1 facts · page: https://llms-explorer.com/tree/mlx-gemv-wide-bf16-small-m-kernel-in-mlx-0-32-1/ ## Facts - Already covered: `gemv_wide` (PR 3888, "metal: add gemv_wide for fp16/bf16 matmuls of a few rows") is listed in the MLX 0.32.1 release notes, and its eligibility, routing and measured speedups are recorded in mlx-gemv-wide-bf16-fp16-few-row-matmul-kernel-pr.md. — [source](https://github.com/ml-explore/mlx/releases/tag/v0.32.1)