Instructions to use InternScience/Agents-A1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use InternScience/Agents-A1 with Transformers:

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

pipe = pipeline("text-generation", model="InternScience/Agents-A1")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("InternScience/Agents-A1")
model = AutoModelForMultimodalLM.from_pretrained("InternScience/Agents-A1", device_map="auto")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.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(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use InternScience/Agents-A1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "InternScience/Agents-A1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "InternScience/Agents-A1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/InternScience/Agents-A1

SGLang

How to use InternScience/Agents-A1 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 "InternScience/Agents-A1" \
    --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": "InternScience/Agents-A1",
		"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 "InternScience/Agents-A1" \
        --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": "InternScience/Agents-A1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use InternScience/Agents-A1 with Docker Model Runner:
```
docker model run hf.co/InternScience/Agents-A1
```

Its better than a lot of models present currently

#30

by testbasis - opened Jul 26

Discussion

testbasis

Jul 26

In my tests Q8 it came out better than the following models with tool use:

Qwen3 Coder Next 8bit

Qwen 3.5 122b A10B HauHau Q5 KM

Qwen 3.6 27B Fable Fus 711 Heretic GGUF Q8_0

Qwen 3.6 27b Abinterated Heretic Uncensored LMX 8 bit

BoZhang

Intern Science org Jul 28

@testbasis

Thank you for sharing your test results — we truly appreciate the feedback and will continue contributing to the community by improving agentic model capabilities across more real-world use cases.

drmcbride

Aug 13

it seems even better than a lot of frontier models for tasks that really require a lot of information to be connected

richard-catcapital

10 days ago

The thread highlights aggregate quality, but agent-runtime behavior often diverges from raw benchmark scores. Specifically, how does this model handle recovery loops after a tool call fails (e.g., returning a 404 or invalid JSON)? Does it degrade into infinite retry cycles, or does it gracefully pivot to alternative tools? Also, does the 'better' performance hold up when the required information span exceeds the context window of standard LLMs, forcing reliance on external retrieval rather than parametric memory?

drmcbride

10 days ago

nice ai slop bro. the whole thing we are saying here is the model was one of the first ones that did a huge number of searches to understand the problem before answering.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment