Instructions to use ibm-granite/granite-4.2-30b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ibm-granite/granite-4.2-30b with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="ibm-granite/granite-4.2-30b")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("ibm-granite/granite-4.2-30b")
model = AutoModelForCausalLM.from_pretrained("ibm-granite/granite-4.2-30b", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
HuggingChat
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use ibm-granite/granite-4.2-30b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ibm-granite/granite-4.2-30b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ibm-granite/granite-4.2-30b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ibm-granite/granite-4.2-30b

SGLang

How to use ibm-granite/granite-4.2-30b with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "ibm-granite/granite-4.2-30b" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ibm-granite/granite-4.2-30b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "ibm-granite/granite-4.2-30b" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ibm-granite/granite-4.2-30b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ibm-granite/granite-4.2-30b with Docker Model Runner:
```
docker model run hf.co/ibm-granite/granite-4.2-30b
```

kswanand1

JustinTong commited on about 20 hours ago

Commit

70fc514

1 Parent(s): a0057a7

Update Serving with SGLang section (#1)

Browse files

- Update Serving with SGLang section (ee753aa1cf9818b8f1a3ff9a40b0c43d0a7b5816)
- Use auto parser resolution and v0.5.18+ in the SGLang section (b5cbf0c68763a412bf52ede7fd39a9c6ff6ab993)

Co-authored-by: Xinyuan Tong <JustinTong@users.noreply.huggingface.co>

Files changed (1) hide show

README.md +10 -11

README.md CHANGED Viewed

@@ -434,15 +434,12 @@ response = client.chat.completions.create(
 print(response.choices[0].message.tool_calls)
 ```
-<!--
 ## Serving with SGLang
-> **Note:** SGLang does not currently support the custom `granite_thinking_parser` natively. This section is commented out until native support is available. In the meantime, use vLLM for serving with the custom reasoning parser.
-Granite-4.2-30B can also be served with [SGLang](https://github.com/sgl-project/sglang) (v0.5.17+) for high-throughput inference.
-> **Reasoning parser:** Use `nemotron_3`.
-> **Tool calling parser:** Use `qwen3_coder`.
 ### Starting the Server
@@ -451,8 +448,8 @@ python3 -m sglang.launch_server \
     --model-path ibm-granite/granite-4.2-30b \
     --dtype bfloat16 \
     --context-length 131072 \
-    --reasoning-parser nemotron_3 \
-    --tool-call-parser qwen3_coder
 ```
 ### OpenAI-Compatible API Usage
@@ -463,13 +460,14 @@ from openai import OpenAI
 client = OpenAI(base_url="http://localhost:30000/v1", api_key="unused")
 response = client.chat.completions.create(
-    model="granite-4.2-30b",
     messages=[{"role": "user", "content": "Explain the Riemann hypothesis in simple terms."}],
     temperature=1.0,
     top_p=0.95,
     max_tokens=8192,
 )
 print(response.choices[0].message.content)
 ```
@@ -498,7 +496,7 @@ tools = [
 ]
 response = client.chat.completions.create(
-    model="granite-4.2-30b",
     messages=[{"role": "user", "content": "What's the weather like in Boston right now?"}],
     tools=tools,
     temperature=1.0,
@@ -508,7 +506,8 @@ response = client.chat.completions.create(
 print(response.choices[0].message.tool_calls)
 ```
--->
 ---

 print(response.choices[0].message.tool_calls)
 ```
 ## Serving with SGLang
+Granite-4.2-30B can also be served with [SGLang](https://github.com/sgl-project/sglang) (v0.5.18+) for high-throughput inference.
+> **Reasoning parser:** Use `--reasoning-parser auto`, which resolves to the built-in `nemotron_3` parser for this checkpoint. It separates the thinking trace into `reasoning_content` and the final answer into `content`, and it handles all three thinking modes (`enable_thinking=True/False`, `low_effort=True`) described in [Thinking Modes](#thinking-modes).
+> **Tool calling parser:** Use `--tool-call-parser auto`, which resolves to `qwen3_coder` for this checkpoint.
 ### Starting the Server
     --model-path ibm-granite/granite-4.2-30b \
     --dtype bfloat16 \
     --context-length 131072 \
+    --reasoning-parser auto \
+    --tool-call-parser auto
 ```
 ### OpenAI-Compatible API Usage
 client = OpenAI(base_url="http://localhost:30000/v1", api_key="unused")
 response = client.chat.completions.create(
+    model="ibm-granite/granite-4.2-30b",
     messages=[{"role": "user", "content": "Explain the Riemann hypothesis in simple terms."}],
     temperature=1.0,
     top_p=0.95,
     max_tokens=8192,
 )
+print(response.choices[0].message.reasoning_content)
 print(response.choices[0].message.content)
 ```
 ]
 response = client.chat.completions.create(
+    model="ibm-granite/granite-4.2-30b",
     messages=[{"role": "user", "content": "What's the weather like in Boston right now?"}],
     tools=tools,
     temperature=1.0,
 print(response.choices[0].message.tool_calls)
 ```
+For a full deployment recipe (Docker, H200/B200 launch matrix, thinking-mode examples, and benchmark data), see the [SGLang Granite 4.2 cookbook](https://docs.sglang.io/cookbook/autoregressive/IBM/Granite-4.2).
 ---