b10375

ggml-org/llama.cppb10375Aug 12, 2026by github-actions[bot]same day as b10373

What changed since b10373

medium risk

b10375 tightens Qwen chat parsing for bare function calls; the imatrix optimization from b10373 is superseded (intermediate b10374 not shown).

New features

  • Stricter bare-function parsing path for Qwen-family chat models, improving tool-call detection reliability.

Improvements

  • Chat template/parser handling for Qwen models made more robust against malformed or ambiguous function-call output.
  • Retains the earlier imatrix performance work that skips finite checks on untouched experts.

Bug fixes

  • Reduces misparsing of bare function syntax in Qwen model responses.

Upgrade notes

No config or CLI changes are required; binaries are drop-in replacements. If you rely on Qwen tool/function calling, re-test prompts that emit bare function syntax, since the parser now rejects looser forms it previously accepted. Note that b10374 was not included in this comparison, so small additional deltas may exist.

Analyzed Oct 31, 2026

Upgrading?

Release notes summary

Refined the chat parser for Qwen models to be more strict regarding bare function parsing to improve compatibility and robustness.

Key Highlights

  • Tightened bare function parsing for Qwen models
  • Improved chat model parsing robustness

Release Notes (excerpt)

<details open> chat : tighten bare function parsing for Qwen models (#26793) </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 7.14) - 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