Instructions to use Lite-Coder/LiteCoder-Terminal-4b-sft with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Lite-Coder/LiteCoder-Terminal-4b-sft with Transformers:

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

pipe = pipeline("text-generation", model="Lite-Coder/LiteCoder-Terminal-4b-sft")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Lite-Coder/LiteCoder-Terminal-4b-sft")
model = AutoModelForCausalLM.from_pretrained("Lite-Coder/LiteCoder-Terminal-4b-sft")
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 Lite-Coder/LiteCoder-Terminal-4b-sft with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Lite-Coder/LiteCoder-Terminal-4b-sft"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Lite-Coder/LiteCoder-Terminal-4b-sft",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Lite-Coder/LiteCoder-Terminal-4b-sft

SGLang

How to use Lite-Coder/LiteCoder-Terminal-4b-sft 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 "Lite-Coder/LiteCoder-Terminal-4b-sft" \
    --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": "Lite-Coder/LiteCoder-Terminal-4b-sft",
		"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 "Lite-Coder/LiteCoder-Terminal-4b-sft" \
        --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": "Lite-Coder/LiteCoder-Terminal-4b-sft",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Lite-Coder/LiteCoder-Terminal-4b-sft with Docker Model Runner:
```
docker model run hf.co/Lite-Coder/LiteCoder-Terminal-4b-sft
```

LiteCoder-Terminal-4b-sft

File size: 4,047 Bytes

---
base_model:
- Qwen/Qwen3-4B-Instruct-2507
license: mit
library_name: transformers
pipeline_tag: text-generation
---

## **LiteCoder-Terminal-4b-sft**

**LiteCoder-Terminal-4b-sft** is part of our latest release on lightweight code agents. The model is fine-tuned from `Qwen3-4B-Instruct-2507` on the [LiteCoder-Terminal-SFT](https://huggingface.co/datasets/Lite-Coder/LiteCoder-Terminal-SFT) dataset. 

This model was introduced in the paper [LiteCoder-Terminal: Scaling Long-Horizon Terminal Environments for Learning Language Agents](https://huggingface.co/papers/2605.29559).

**Code:** [https://github.com/icip-cas/LiteCoder](https://github.com/icip-cas/LiteCoder)

Compared to our previous preview version, we scaled up the training data from under 1,000 samples to 11,255 trajectories, incorporating a broader task taxonomy and diverse agent scaffolds. With these updates, the model shows consistent improvements across Terminal Bench evaluations.

## **Released Artifacts**

| 2026/04/13 |  |  |
| --- | --- | --- |
| LiteCoder-Terminal-30b-a3b-sft | Model | [**https://huggingface.co/Lite-Coder/LiteCoder-Terminal-30b-a3b-sft**](https://huggingface.co/Lite-Coder/LiteCoder-Terminal-30b-a3b-sft) |
| LiteCoder-Terminal-4b-sft | Model | [**https://huggingface.co/Lite-Coder/LiteCoder-Terminal-4b-sft**](https://huggingface.co/Lite-Coder/LiteCoder-Terminal-4b-sft) |
| LiteCoder-Terminal-SFT | Dataset | [**https://huggingface.co/datasets/Lite-Coder/LiteCoder-Terminal-SFT**](https://huggingface.co/datasets/Lite-Coder/LiteCoder-Terminal-SFT/settings) |
| LiteCoder-Terminal-World-Model-SFT | Dataset | [**https://huggingface.co/datasets/Lite-Coder/LiteCoder-Terminal-World-Model-SFT**](https://huggingface.co/datasets/Lite-Coder/LiteCoder-Terminal-World-Model-SFT) |
| LiteCoder-Terminal-RL-preview | Dataset | [**https://huggingface.co/datasets/Lite-Coder/LiteCoder-Terminal-RL-preview**](https://huggingface.co/datasets/Lite-Coder/LiteCoder-Terminal-RL-preview/tree/main) |

## **Results**

### Terminal Bench 1.0 Performance

| Model | Agent | pass@1 | pass@4 |
| --- | --- | --- | --- |
| **LiteCoder-Terminal-30b-a3b-sft** | Terminus 2 | **24.38%** | **40%** |
| Qwen3-30B-A3B-Nex-N1 | Openhands | 18.44% | 32.5% |
| LiteCoder-30a3b-Terminal-preview | Terminus 2 | 16.56% | 27.5% |
| Qwen3-30B-A3B-Instruct | Terminus 2 | 16.56% | 28.75% |
| **LiteCoder-Terminal-4b-sft** | Terminus 2 | **13.44%** | **30%** |
| OpenThinker-Agent-v1 | Terminus 2 | 11.25% | 25% |
| LiteCoder-4b-Terminal-preview | Terminus 2 | 9.38% | 20% |
| Qwen3-4B-Instruct | Terminus 2 | 6.25% | 15% |

### Terminal Bench 2.0 Performance

| Model | Agent | pass@1 | pass@4 |
| --- | --- | --- | --- |
| **LiteCoder-Terminal-30b-a3b-sft** | Terminus 2 | **12.36%** | **23.60%** |
| Qwen3-30B-A3B-Nex-N1 | Openhands | 12.36% | 23.60% |
| LiteCoder-30a3b-Terminal-preview | Terminus 2 | 6.18% | 13.75% |
| **LiteCoder-Terminal-4b-sft** | Terminus 2 | **5.62%** | **12.36%** |
| Qwen3-30B-A3B-Instruct | Terminus 2 | 5.34% | 11.24% |
| OpenThinker-Agent-v1 | Terminus 2 | 4.49% | 10.1% |
| LiteCoder-4b-Terminal-preview | Terminus 2 | 4.78% | 12.36% |
| Qwen3-4B-Instruct | Terminus 2 | 1.12% | 3.37% |

### Terminal Bench Pro Performance

| Model | Agent | pass@1 |
| --- | --- | --- |
| **LiteCoder-Terminal-30b-a3b-sft** | Terminus 2 | **31.5%** |
| LiteCoder-30a3b-Terminal-preview | Terminus 2 | 22.0% |
| Qwen3-30B-A3B-Nex-N1 | Openhands | 21.0% |
| Qwen3-30B-A3B-Instruct | Terminus 2 | 20.5% |
| OpenThinker-Agent-v1 | Terminus 2 | 19.5% |
| **LiteCoder-Terminal-4b-sft** | Terminus 2 | **15.5%** |
| LiteCoder-4b-Terminal-preview | Terminus 2 | 15.0% |
| Qwen3-4B-Instruct | Terminus 2 | 3.5% |

## **Citation**

```bash
@article{peng2026litecoderterminal,
  title={LiteCoder-Terminal: Scaling Long-Horizon Terminal Environments for Learning Language Agents},
  author={Peng, Xiaoxuan and Zhang, Kaiqi and Lu, Xinyu and Cao, Boxi and Lu, Yaojie and Lin, Hongyu and Han, Xianpei and Sun, Le},
  journal={arXiv preprint arXiv:2605.29559},
  year={2026}
}
```