Instructions to use Salesforce/xgen-small-9B-base-r with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Salesforce/xgen-small-9B-base-r with Transformers:

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

pipe = pipeline("text-generation", model="Salesforce/xgen-small-9B-base-r")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Salesforce/xgen-small-9B-base-r")
model = AutoModelForCausalLM.from_pretrained("Salesforce/xgen-small-9B-base-r")
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]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Salesforce/xgen-small-9B-base-r with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Salesforce/xgen-small-9B-base-r"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Salesforce/xgen-small-9B-base-r",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Salesforce/xgen-small-9B-base-r

SGLang

How to use Salesforce/xgen-small-9B-base-r 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 "Salesforce/xgen-small-9B-base-r" \
    --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": "Salesforce/xgen-small-9B-base-r",
		"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 "Salesforce/xgen-small-9B-base-r" \
        --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": "Salesforce/xgen-small-9B-base-r",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Salesforce/xgen-small-9B-base-r with Docker Model Runner:
```
docker model run hf.co/Salesforce/xgen-small-9B-base-r
```

Welcome to the xGen-small family!

xGen-small (blog, arXiv) is an enterprise-ready compact LM that combines domain-focused data curation, scalable pre-training, length-extension, and RL fine-tuning to deliver long-context performance at predictable, low cost. This model release is for research purposes only.

Model Series

xGen-small comes in two sizes (4B and 9B) with two variants (pre-trained and post-trained):

Model	# Total Params	Context Length	Variant	Download
salesforce/xgen-small-4B-base-r	4B	128k	Pre-trained	🤗 Link
salesforce/xgen-small-4B-instruct-r	4B	128k	Post-trained	🤗 Link
salesforce/xgen-small-9B-base-r	9B	128k	Pre-trained	🤗 Link
salesforce/xgen-small-9B-instruct-r	9B	128k	Post-trained	🤗 Link

Usage

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "Salesforce/xgen-small-9B-base-r"
tokenizer = AutoTokenizer.from_pretrained(model_name)
device = "cuda" if torch.cuda.is_available() else "cpu"
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype="auto"
).to(device)

prompt = "What is Salesforce?"
inputs = tokenizer(
    prompt,
    return_tensors="pt",
    padding=False,
    truncation=True
).to(device)

generated = model.generate(**inputs, max_new_tokens=32)
output = tokenizer.decode(
    generated[0],
    skip_special_tokens=True,
)
print(output)

Evaluation

Category	Task	Llama 3.1-8B	Granite 3.3-8B	Qwen2.5-7B	xGen-small 9B Base
General Knowledge & Reasoning	ARC-Challenge	58.0	62.5	63.7	67.4
General Knowledge & Reasoning	Big-Bench Hard	46.3	46.8	53.6	58.2
General Knowledge & Reasoning	HellaSwag	81.8	83.0	80.0	83.7
General Knowledge & Reasoning	MMLU	65.1	62.7	74.2	71.1
General Knowledge & Reasoning	MMLU-Pro	32.7	31.3	43.7	39.8
General Knowledge & Reasoning	TruthfulQA	45.2	52.2	56.4	48.6
General Knowledge & Reasoning	WinoGrande	76.9	80.3	76.1	78.6
Math & Science	GPQA	31.9	30.3	31.4	32.0
Math & Science	GSM8K	55.6	61.4	79.1	83.2
Math & Science	MATH	22.0	30.9	50.2	52.5
Coding	HumanEval	37.3	38.9	55.2	53.9
Coding	HumanEval+	31.4	34.3	47.7	47.9
Coding	MBPP	45.0	43.5	57.1	50.1
Coding	MBPP+	51.0	48.1	64.8	57.6

Citation

@misc{xgensmall,
      title={xGen-small Technical Report}, 
      author={Erik Nijkamp and Bo Pang and Egor Pakhomov and Akash Gokul and Jin Qu and Silvio Savarese and Yingbo Zhou and Caiming Xiong},
      year={2025},
      eprint={2505.06496},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2505.06496}, 
}

Ethical Considerations

This release is for research purposes only in support of an academic paper. Our models, datasets, and code are not specifically designed or evaluated for all downstream purposes. We strongly recommend users evaluate and address potential concerns related to accuracy, safety, and fairness before deploying this model. We encourage users to consider the common limitations of AI, comply with applicable laws, and leverage best practices when selecting use cases, particularly for high-risk scenarios where errors or misuse could significantly impact people's lives, rights, or safety. For further guidance on use cases, refer to our AUP and AI AUP.

Model Licenses

Downloads last month: 39

Safetensors

Model size

11B params

Tensor type

F32

Paper for Salesforce/xgen-small-9B-base-r

xGen-small Technical Report

Paper • 2505.06496 • Published May 10, 2025