b10288

ggml-org/llama.cppb10288Aug 5, 2026by github-actions[bot]same day as b10287

What changed since b10287

low risk

Test-only release: MiniMax M3 re-enabled in the architecture test suite, with no runtime or API changes.

Improvements

  • MiniMax M3 architecture is once again covered by the test-llama-archs suite, restoring regression coverage for that model family.
  • Prebuilt binaries refreshed across all usual platforms and backends (macOS, iOS, Linux CPU/Vulkan/ROCm/OpenVINO/SYCL, Windows CUDA/HIP, Android).

Upgrade notes

This build only touches test infrastructure, so no config, model, or API migration is required. Users who do not run the project's own test suite will see no behavioral difference from b10287. The Unlimited-OCR max_tiles converter fix from the previous build remains included.

Analyzed Oct 22, 2026

Upgrading?

Release notes summary

This build focuses on test infrastructure improvements, specifically re-enabling the MiniMax M3 architecture test. It includes standard pre-built binaries for multiple platforms and operating systems.

Key Highlights

  • Re-enabled MiniMax M3 architecture in test suite.
  • Available binaries for macOS (Apple Silicon/Intel) and iOS.
  • Available binaries for Linux (various backends: Vulkan, ROCm, OpenVINO, SYCL).
  • Available binaries for Windows (CPU, CUDA 12/13, Vulkan, OpenVINO, SYCL, HIP).
  • Available Android binary.

Release Notes (excerpt)

<details open> tests: re-enable MiniMax M3 in `test-llama-archs` (#26633) </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