llama-server & OpenAI endpoint

Deploying via llama-server with an OpenAI compatible endpoint

We are doing to deploy Devstral-2 - see Devstral 2 for more details on the model.

Obtain the latest llama.cpp on GitHub here. You can follow the build instructions below as well. Change -DGGML_CUDA=ON to -DGGML_CUDA=OFF if you don't have a GPU or just want CPU inference.

apt-get update
apt-get install pciutils build-essential cmake curl libcurl4-openssl-dev -y
git clone https://github.com/ggml-org/llama.cpp
cmake llama.cpp -B llama.cpp/build \
    -DBUILD_SHARED_LIBS=OFF -DGGML_CUDA=ON -DLLAMA_CURL=ON
cmake --build llama.cpp/build --config Release -j --clean-first --target llama-cli llama-mtmd-cli llama-server llama-gguf-split
cp llama.cpp/build/bin/llama-* llama.cpp

First download Devstral 2:

# !pip install huggingface_hub hf_transfer
import os
os.environ["HF_HUB_ENABLE_HF_TRANSFER"] = "1"
from huggingface_hub import snapshot_download
snapshot_download(
    repo_id = "unsloth/Devstral-2-123B-Instruct-2512-GGUF",
    local_dir = "Devstral-2-123B-Instruct-2512-GGUF",
    allow_patterns = ["*UD-Q2_K_XL*", "*mmproj-F16*"],
)

To deploy Devstral 2 for production, we use llama-server In a new terminal say via tmux, deploy the model via:

./llama.cpp/llama-server \
    --model Devstral-Small-2-24B-Instruct-2512-GGUF/Devstral-Small-2-24B-Instruct-2512-UD-Q4_K_XL.gguf \
    --mmproj Devstral-Small-2-24B-Instruct-2512-GGUF/mmproj-F16.gguf \
    --alias "unsloth/Devstral-Small-2-24B-Instruct-2512" \
    --threads -1 \
    --n-gpu-layers 999 \
    --prio 3 \
    --min_p 0.01 \
    --ctx-size 16384 \
    --port 8001 \
    --jinja

When you run the above, you will get:

Then in a new terminal, after doing pip install openai, do:

from openai import OpenAI
import json
openai_client = OpenAI(
    base_url = "http://127.0.0.1:8001/v1",
    api_key = "sk-no-key-required",
)
completion = openai_client.chat.completions.create(
    model = "unsloth/Devstral-Small-2-24B-Instruct-2512",
    messages = [{"role": "user", "content": "What is 2+2?"},],
)
print(completion.choices[0].message.content)

Which will simply print 4. You can go back to the llama-server screen and you might see some statistics which might be interesting:

For arguments like using speculative decoding, see https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md