b10198

ggml-org/llama.cppb10198Jul 30, 2026by github-actions[bot]same day as b10197

What changed since b10197

low risk

b10198 adds Vulkan backend support for quantized concat operations, with no other changes over b10197.

New features

  • Vulkan backend can now execute quantized concat operations, which previously fell back or were unsupported.

Upgrade notes

This is an incremental build with a single Vulkan backend addition; no config, model, or CLI changes are involved. Vulkan users running models that hit quantized concat ops should see those graphs stay on-device instead of falling back. Rebuild or pull the new Vulkan binaries to pick up the change.

Analyzed Oct 31, 2026

Upgrading?

Release notes summary

This release expands Vulkan backend capabilities by adding support for quantized concatenation operations.

Key Highlights

  • Added support for quantized concat operations in the Vulkan backend

Release Notes (excerpt)

<details open> vulkan: Support quantized concat (#25684) </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:** - Android arm64 (CPU)…

Published by ggml-org on GitHub.

Read the full release notes

Related Repositories