b10234

ggml-org/llama.cppb10234Aug 2, 2026by github-actions[bot]same day as b10233

What changed since b10233

low risk

b10234 adds F16 support for Metal binary operations, following b10233's OpenCL GLU workgroup fix.

New features

  • Metal backend now supports F16 (half-precision) binary operations.

Improvements

  • Half-precision models running on Metal can execute binary ops natively, improving throughput.
  • Broadens Metal backend kernel coverage for F16 workloads.

Upgrade notes

No configuration or API changes are required; simply pull the b10234 binaries or rebuild with Metal enabled to pick up the new F16 kernels. The OpenCL GLU workgroup limit from b10233 remains in place, so OpenCL users see no regression. Apple Silicon users are the primary beneficiaries of this delta.

Analyzed Oct 31, 2026

Upgrading?

Release notes summary

Added F16 (half-precision floating point) support for binary operations in the Metal backend.

Key Highlights

  • Added F16 support for binary operations.
  • Improves performance for models utilizing half-precision arithmetic.
  • Expands Metal backend capabilities.

Release Notes (excerpt)

<details open> metal : add F16 support for bin ops (#26465) </details> **Website:** - <https://llama.app> **macOS/iOS:** - macOS Apple Silicon (arm64) - macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED - macOS Intel (x64) - iOS XCFramework **Linux:** - Ubuntu x64 (CPU) - Ubuntu arm64 (CPU) - Ubuntu s390x (CPU) - Ubuntu x64 (Vulkan) - Ubuntu arm64 (Vulkan) - Ubuntu x64 (ROCm 7.2) - Ubuntu x64 (OpenVINO) - Ubuntu x64 (SYCL FP32) - Ubuntu x64 (SYCL FP16) **Android:**…

Published by ggml-org on GitHub.

Read the full release notes

Related Repositories