b9630

ggml-org/llama.cppb9630Jun 14, 2026by github-actions[bot]same day as b9628

What changed since b9628

low risk

Adds Cohere2-MoE support to the TINY_AYA tokenizer vocabulary; otherwise a routine rebuild.

New features

  • TINY_AYA tokenizer vocabulary now recognizes the cohere2moe architecture.

Improvements

  • Refreshed prebuilt binaries for all platforms at the new build number.

Upgrade notes

This is a narrow tokenizer-vocabulary addition with no config or API changes; existing models and flags continue to work. Users working with Cohere2-MoE models via the TINY_AYA vocab should pick up this build. Otherwise the upgrade is a drop-in binary swap.

Analyzed Oct 31, 2026

Upgrading?

Release notes summary

Updated the TINY_AYA tokenizer vocabulary to include support for the Cohere2-MoE architecture.

Key Highlights

  • Added cohere2moe to llama-vocab for TINY_AYA
  • Updated binaries for all platforms

Release Notes (excerpt)

<details open> Add cohere2moe to llama-vocab for TINY_AYA (#24601) </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:** - Android arm64 (CPU) **Windows:**…

Published by ggml-org on GitHub.

Read the full release notes

Related Repositories