Add llama.cpp PR build instructions

README.md

@@ -57,6 +57,30 @@ Note: MiniMax Sparse Attention is not supported yet, so inference falls back to
 
 # MiniMax-M3
 
+## Run MiniMax-M3 in llama.cpp
+
+MiniMax-M3 support in llama.cpp is preliminary and not yet in a released build. To run these GGUFs, build llama.cpp from [PR #24523](https://github.com/ggml-org/llama.cpp/pull/24523):
+
+```bash
+git clone https://github.com/ggml-org/llama.cpp
+cd llama.cpp
+git fetch origin pull/24523/head:minimax-m3
+git checkout minimax-m3
+cmake -B build -DGGML_CUDA=ON
+cmake --build build --config Release -j --target llama-cli llama-server
+```
+
+Then run a quant. The model is large (~428B params), so offload across GPUs with `-ngl 99` or keep the weights in CPU RAM:
+
+```bash
+./build/bin/llama-cli \
+  -hf unsloth/MiniMax-M3-GGUF:UD-Q4_K_XL \
+  --jinja -ngl 99 --ctx-size 8192 \
+  -p "Hello, who are you?"
+```
+
+Note: MiniMax Sparse Attention is not supported yet, so inference falls back to dense attention.
+
 MiniMax-M3 is a native multimodal model with 1M context. It has ~428B parameters and ~23B activated parameters.
 
 **Highlights:**