b9940

ggml-org/llama.cppb9940Jul 9, 2026by github-actions[bot]same day as b9939

What changed since b9939

low risk

b9940 adds an offline mode to llama-bench, letting benchmarks run without network access.

New features

  • llama-bench now initializes params.offline, enabling benchmarking in network-isolated environments.

Improvements

  • Benchmark runs become more reproducible when no network is available.

Upgrade notes

This is a single-commit increment over b9939 with no config or API changes required. Users of llama-bench in air-gapped setups can now run benchmarks without network access; no migration steps are needed.

Analyzed Oct 24, 2026

Upgrading?

Release notes summary

Added offline parameter support to the llama-bench benchmarking tool.

Key Highlights

  • Added init params.offline for llama-bench
  • Allows benchmarking without requiring network access
  • Improves benchmark reproducibility in offline environments

Release Notes (excerpt)

<details open> llama-bench : init params.offline (#25476) Signed-off-by: Adrien Gallouët <angt@huggingface.co> </details> **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