Instructions to use SearchSwarm/SearchSwarm-30B-A3B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use SearchSwarm/SearchSwarm-30B-A3B with Transformers:

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

pipe = pipeline("text-generation", model="SearchSwarm/SearchSwarm-30B-A3B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForMultimodalLM

tokenizer = AutoTokenizer.from_pretrained("SearchSwarm/SearchSwarm-30B-A3B")
model = AutoModelForMultimodalLM.from_pretrained("SearchSwarm/SearchSwarm-30B-A3B")
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 Settings

vLLM

How to use SearchSwarm/SearchSwarm-30B-A3B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "SearchSwarm/SearchSwarm-30B-A3B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "SearchSwarm/SearchSwarm-30B-A3B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/SearchSwarm/SearchSwarm-30B-A3B

SGLang

How to use SearchSwarm/SearchSwarm-30B-A3B 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 "SearchSwarm/SearchSwarm-30B-A3B" \
    --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": "SearchSwarm/SearchSwarm-30B-A3B",
		"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 "SearchSwarm/SearchSwarm-30B-A3B" \
        --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": "SearchSwarm/SearchSwarm-30B-A3B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use SearchSwarm/SearchSwarm-30B-A3B with Docker Model Runner:
```
docker model run hf.co/SearchSwarm/SearchSwarm-30B-A3B
```

SearchSwarm-30B-A3B

SearchSwarm: Towards Delegation Intelligence in Agentic LLMs for Long-Horizon Deep Research

Project Page | GitHub Repository

SearchSwarm is a foundation action model designed for long-horizon deep research tasks. It introduces delegation intelligence, enabling a main research agent to use subagents as an active context-management mechanism. The main agent decomposes complex tasks, dispatches bounded subtasks to subagents, and synthesizes citation-grounded reports while staying within a finite context budget.

Overview

SearchSwarm focuses on several key aspects of agentic intelligence:

Subagents as Context Management: Subagents work in independent contexts and return compact, evidence-grounded reports, preventing the main agent's context window from being overwhelmed.
Delegation Intelligence: The model is trained to determine when and what to delegate, how to brief subagents effectively, and how to verify their findings.
Harness-guided Synthesis: The authors used a specialized harness to synthesize high-quality SFT data to internalize these delegation capabilities into the model weights.

Performance

SearchSwarm-30B-A3B achieves state-of-the-art results among comparable 30B-A3B open-source lightweight research agents on deep research benchmarks:

Benchmark	Score
BrowseComp	68.1
BrowseComp-ZH	73.3

Citation

@misc{searchswarm2026,
  title        = {SearchSwarm: Towards Delegation Intelligence in Agentic LLMs for Long-Horizon Deep Research},
  author       = {Ning, Pu and Chen, Quan and Tao, Kun and Tang, Xinyu and Wang, Tianshu and Cao, Qianggang and Kong, Xinyu and Wen, Zujie and Zhang, Zhiqiang and Zhou, Jun},
  year         = {2026},
  note         = {Under review}
}

Acknowledgements

This project builds on open-source infrastructure including vLLM, ms-swift, Megatron-LM, Qwen-Agent, Serper, and Jina.

Downloads last month: 24

Safetensors

Model size

31B params

Tensor type

BF16

Model tree for SearchSwarm/SearchSwarm-30B-A3B

Base model

Alibaba-NLP/Tongyi-DeepResearch-30B-A3B

Finetuned

(3)

this model

Paper for SearchSwarm/SearchSwarm-30B-A3B

SearchSwarm: Towards Delegation Intelligence in Agentic LLMs for Long-Horizon Deep Research

Paper • 2606.09730 • Published 4 days ago • 49