Llama.cpp Expands Support for Latest AI Models
This saga tracks the rapid integration of new model architectures and decoding optimizations into the llama.cpp inference engine. The latest update adds support for DeepSeek-V4.1-Flash, continuing a trend of expanding compatibility with emerging AI systems.
-
Llama.cpp adds support for xing4_0 model
A pull request merged into the llama.cpp repository on November 3, 2026, added support for the xing40 model. The change was co-authored by Sigbjørn Skjæret of Hugging Face and reviewed internally by…
1 source primary source -
Ggml-org adds Clover model support to llama.cpp
The open-source project llama.cpp now supports the Clover model, announced via a pull request merged on October 28, 2026. The update includes GGUF conversion, model loading, and causal hybrid…
1 source primary source -
Llama.cpp PR adds DeepSeek-V4.1-Flash support with review fixes
A pull request for the llama.cpp repository introduces support for the DeepSeek-V4.1-Flash model architecture, including its specific KV-cache extensions, vision components, and DSpark draft…
1 source primary source -
Ggml-org merges webgpu Q2_0 support into llama.cpp
A pull request (PR 30914) adding Q20 quantization support for WebGPU was merged into the ggml‑org/llama.cpp repository. The change, labeled “merge ready” by maintainer yomaytk on Oct 21‑22, 2026,…
2 sources primary source -
Ggml-org adds fused top‑k mask operation to llama.cpp
A pull request to the ggml‑org/llama.cpp repository introduces a new fused operation for QSA masking. The change adds the GGMLOPTOPKMASK node to the CUDA backend, allowing the mask to be applied…
1 source primary source -
Ggml-org adds `--spec-draft-n-ubatch` to llama.cpp for batch control
llama.cpp now lets users set a custom physical batch size for draft context via the --spec-draft-n-ubatch flag. The change, merged in a pull request by ggml-org, allows drafts to process tokens in…
1 source primary source -
Llama.cpp adds MTP decoding for Qwen4Exp and GLM-5.3-Flash hybrid model
llama.cpp version 0.6.0 introduces MTP (Mixture of Token Predictors) speculative decoding for the Qwen4Exp model, claiming a ~1.5x speedup on DGX Spark hardware. The update also adds support for the…
2 sources primary source -
Qwen4Exp adds MTP support in llama.cpp pull request
A pull request merged into the llama.cpp repository on October 1, 2026, adds support for MTP (likely referring to Mixture of Token Predictions or a similar technique) to the Qwen4Exp model. The…
2 sources primary source