Models & Labs
llama.cpp b10085 release updates Qwen3-VL interpolation
llama.cpp Releaseshigh confidence
Why it matters
- →Aligns interpolation method with transformers reference for accuracy.
- →Reduces discrepancies in image scaling, crucial for AI tasks.
- →Enhances reliability of llama.cpp for developers.
The b10085 release of llama.cpp focuses on improving the Qwen3-VL vision model's position embedding interpolation. The update aligns the interpolation method with the transformers reference, using align_corners=True, which corrects scaling issues for grounding coordinates in images. This adjustment is particularly important for larger and non-square images, ensuring more accurate image processing. The release does not introduce new models but enhances the precision of existing features, making it a valuable update for developers.
Read originalThe story around this
Earlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
How we got here
- 1llama.cpp b9684 Release Adds 3D Convolution
llama.cpp Releases · June 18, 2026 · Same story
- 2llama.cpp b9850 Release Enhances Qwen3 Model
llama.cpp Releases · July 2, 2026 · Same story
More from llama.cpp Releases
Coding Toolscoding
llama.cpp b12022 adds ROCm 10 and Snapdragon support
This release quietly cements llama.cpp as the universal inference runtime by finally bringing ROCm 10.0 to Linux and Windows alongside CUDA 13.4, effectively closing the hardware gap for AMD users who were previously left behind. The inclusion of native Snapdragon support on Linux marks a significant step toward running local LLMs on ARM-based mobile chips without emulation overhead. While KleidiAI on Apple Silicon is disabled in this build, the broader expansion to diverse accelerators ensures that developers aren't locked into NVIDIA ecosystems anymore.
llama.cpp Releases
Coding Toolscoding
llama.cpp b12023 fixes shared prefix bug
This release addresses a specific inference artifact where decision prefixes were incorrectly bleeding into output embeddings. By capping shared prefixes at the first output position, llama.cpp ensures that option tokens remain distinct for parent and child tasks. This is a targeted fix for multi-step reasoning or voting workflows rather than a performance boost. It stabilizes complex generation patterns without changing the underlying model architecture.
