File size: 3,602 Bytes
33db589 40ff83a | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 | ---
license: mit
base_model:
- Qwen/Qwen3-4B-Instruct-2507
---
## **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.
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%** | **30%** |
| 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
@misc{litecoder2026,
title={LiteCoder: Advancing Small and Medium-sized Code Agents},
author={Xiaoxuan Peng and Xinyu Lu and Kaiqi Zhang and Taosong Fang and Boxi Cao and Yaojie Lu},
year={2026},
}
``` |