Instructions to use OrionLLM/OxCoder-9B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OrionLLM/OxCoder-9B with Transformers:

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

pipe = pipeline("text-generation", model="OrionLLM/OxCoder-9B")
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("OrionLLM/OxCoder-9B")
model = AutoModelForMultimodalLM.from_pretrained("OrionLLM/OxCoder-9B", 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 OrionLLM/OxCoder-9B with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/OrionLLM/OxCoder-9B

SGLang

How to use OrionLLM/OxCoder-9B 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 "OrionLLM/OxCoder-9B" \
    --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": "OrionLLM/OxCoder-9B",
		"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 "OrionLLM/OxCoder-9B" \
        --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": "OrionLLM/OxCoder-9B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use OrionLLM/OxCoder-9B with Docker Model Runner:
```
docker model run hf.co/OrionLLM/OxCoder-9B
```

OxCoder-9B / README.md

DedeProGames

Update README.md

a1d6528 verified about 12 hours ago

preview code

Raw

History Blame Contribute Delete

13.2 kB

	---
	library_name: transformers
	license: apache-2.0
	pipeline_tag: text-generation
	base_model:
	- Qwen/Qwen3.5-9B
	---

	![Banner](https://cdn-uploads.huggingface.co/production/uploads/685ea8ff7b4139b6845ce395/vn7pEOMUQ9-5vYgJQ4oVh.png)

	## Introduction

	We are proud to introduce OxCoder-9B, our lightweight coding model for long-horizon tasks, agentic coding, and agentic reasoning. Despite featuring only 9 billion parameters, OxCoder-9B marks a substantial leap in agentic capabilities — particularly in Agentic Terminal and Agentic Coding tasks — punching far above its weight class and rivaling models many times its size. This release represents our strongest commitment yet to delivering frontier-level agentic performance in a compact, efficient, and openly accessible package.

	- Trained on Frontier Agent Traces: Built from Fable-5.1 and GLM-5.3 agentic coding trajectories across Claude Code, OpenCode and Codex — OxCoder-9B has been distilled from some of the most capable agentic systems available, giving it a unique advantage in real-world software engineering scenarios.
	- 262K Native Context: Equipped with a full 262,144 token native context window, enabling the model to handle complex, multi-file codebases and long-horizon reasoning tasks with ease.
	- Error Recovery: Learns read-before-write patterns, responds to LSP diagnostics, and applies minimal edit diffs instead of full rewrites — making it robust in iterative coding environments where precision matters.
	- Front-end Reasoning: OxCoder-9B exhibits remarkably strong front-end reasoning capabilities for its size, demonstrating a deep understanding of UI logic, component architecture, and web-native patterns that is rare in sub-10B models.

	![Coding Performance Evaluation](./assets/coding-performance-evaluation.png)


	## Benchmark

	<table>
	<tr>
	<th style="background: rgba(128,128,128,0.1); text-align: center; border: hidden;"> </th>
	<th style="background: rgba(128,128,128,0.1); text-align: center; border: hidden;">OxCoder-9B</th>
	<th style="background: rgba(128,128,128,0.1); text-align: center; border: hidden;">Ornith-1.5-9B</th>
	<th style="background: rgba(128,128,128,0.1); text-align: center; border: hidden;">Ornith-1.0-9B</th>
	<th style="background: rgba(128,128,128,0.1); text-align: center; border: hidden;">Qwen3.5-9B</th>
	<th style="background: rgba(128,128,128,0.1); text-align: center; border: hidden;">Gemma-4-31B</th>
	</tr>
	<tr>
	<td align="center" colspan="6" style="background: linear-gradient(90deg, rgba(124,58,237,0.45) 0%, rgba(99,102,241,0.42) 50%, rgba(59,130,246,0.45) 100%); font-weight: bold; height:32px; padding-top:2px; padding-bottom:2px;"><i>Coding</i></td>
	</tr>
	<tr>
	<td style="text-align:left; padding:9px 10px 9px 18px;"><div style="font-size:15px; font-weight:600; line-height:1.22;">Agentic terminal coding</div><div style="margin-top:4px; font-size:11px; font-weight:400; line-height:1.2; opacity:0.62;">Terminal-Bench 2.1 (Terminus-2)</div></td>
	<td align="center" style="background: linear-gradient(135deg, rgba(124,58,237,0.16) 0%, rgba(118,65,238,0.16) 100%); vertical-align:middle;"><b>49.6</b></td>
	<td align="center" style="vertical-align:middle;">46.2</td>
	<td align="center" style="vertical-align:middle;">43.1</td>
	<td align="center" style="vertical-align:middle;">21.3</td>
	<td align="center" style="vertical-align:middle;">42.1</td>
	</tr>
	<tr>
	<td style="text-align:left; padding:9px 10px 9px 18px;"><div style="font-size:15px; font-weight:600; line-height:1.22;">Agentic terminal coding</div><div style="margin-top:4px; font-size:11px; font-weight:400; line-height:1.2; opacity:0.62;">Terminal-Bench 2.1 (Claude Code)</div></td>
	<td align="center" style="background: linear-gradient(135deg, rgba(118,65,238,0.16) 0%, rgba(112,71,239,0.16) 100%); vertical-align:middle;"><b>50.8</b></td>
	<td align="center" style="vertical-align:middle;">47.0</td>
	<td align="center" style="vertical-align:middle;">40.6</td>
	<td align="center" style="vertical-align:middle;">18.9</td>
	<td align="center" style="vertical-align:middle;">—</td>
	</tr>
	<tr>
	<td style="text-align:left; padding:9px 10px 9px 18px;"><div style="font-size:15px; font-weight:600; line-height:1.22;">Agentic coding</div><div style="margin-top:4px; font-size:11px; font-weight:400; line-height:1.2; opacity:0.62;">SWE-bench Verified</div></td>
	<td align="center" style="background: linear-gradient(135deg, rgba(112,71,239,0.16) 0%, rgba(106,78,239,0.16) 100%); vertical-align:middle;"><b>73.5</b></td>
	<td align="center" style="vertical-align:middle;">70.6</td>
	<td align="center" style="vertical-align:middle;">69.4</td>
	<td align="center" style="vertical-align:middle;">53.2</td>
	<td align="center" style="vertical-align:middle;">52.0</td>
	</tr>
	<tr>
	<td style="text-align:left; padding:9px 10px 9px 18px;"><div style="font-size:15px; font-weight:600; line-height:1.22;">Real-world software engineering</div><div style="margin-top:4px; font-size:11px; font-weight:400; line-height:1.2; opacity:0.62;">SWE-bench Pro</div></td>
	<td align="center" style="background: linear-gradient(135deg, rgba(106,78,239,0.16) 0%, rgba(100,84,240,0.16) 100%); vertical-align:middle;"><b>49.1</b></td>
	<td align="center" style="vertical-align:middle;">47.5</td>
	<td align="center" style="vertical-align:middle;">42.9</td>
	<td align="center" style="vertical-align:middle;">31.3</td>
	<td align="center" style="vertical-align:middle;">35.7</td>
	</tr>
	<tr>
	<td style="text-align:left; padding:9px 10px 9px 18px;"><div style="font-size:15px; font-weight:600; line-height:1.22;">Repo-level code generation</div><div style="margin-top:4px; font-size:11px; font-weight:400; line-height:1.2; opacity:0.62;">NL2Repo</div></td>
	<td align="center" style="background: linear-gradient(135deg, rgba(100,84,240,0.16) 0%, rgba(94,91,241,0.16) 100%); vertical-align:middle;"><b>36.2</b></td>
	<td align="center" style="vertical-align:middle;">32.4</td>
	<td align="center" style="vertical-align:middle;">27.2</td>
	<td align="center" style="vertical-align:middle;">16.2</td>
	<td align="center" style="vertical-align:middle;">15.5</td>
	</tr>
	<tr>
	<td align="center" colspan="6" style="background: linear-gradient(90deg, rgba(124,58,237,0.45) 0%, rgba(99,102,241,0.42) 50%, rgba(59,130,246,0.45) 100%); font-weight: bold; height:32px; padding-top:2px; padding-bottom:2px;"><i>Reasoning</i></td>
	</tr>
	<tr>
	<td style="text-align:left; padding:9px 10px 9px 18px;"><div style="font-size:15px; font-weight:600; line-height:1.22;">Expert-level reasoning</div><div style="margin-top:4px; font-size:11px; font-weight:400; line-height:1.2; opacity:0.62;">HLE (no tools)</div></td>
	<td align="center" style="background: linear-gradient(135deg, rgba(94,91,241,0.16) 0%, rgba(89,97,242,0.16) 100%); vertical-align:middle;"><b>21.2</b></td>
	<td align="center" style="vertical-align:middle;">20.2</td>
	<td align="center" style="vertical-align:middle;">16.8</td>
	<td align="center" style="vertical-align:middle;">14.7</td>
	<td align="center" style="vertical-align:middle;">19.5</td>
	</tr>
	<tr>
	<td style="text-align:left; padding:9px 10px 9px 18px;"><div style="font-size:15px; font-weight:600; line-height:1.22;">Tool-augmented reasoning</div><div style="margin-top:4px; font-size:11px; font-weight:400; line-height:1.2; opacity:0.62;">HLE (with tools)</div></td>
	<td align="center" style="background: linear-gradient(135deg, rgba(89,97,242,0.16) 0%, rgba(83,104,243,0.16) 100%); vertical-align:middle;"><b>32.8</b></td>
	<td align="center" style="vertical-align:middle;">30.5</td>
	<td align="center" style="vertical-align:middle;">26.4</td>
	<td align="center" style="vertical-align:middle;">24.5</td>
	<td align="center" style="vertical-align:middle;">26.5</td>
	</tr>
	<tr>
	<td style="text-align:left; padding:9px 10px 9px 18px;"><div style="font-size:15px; font-weight:600; line-height:1.22;">Scientific reasoning</div><div style="margin-top:4px; font-size:11px; font-weight:400; line-height:1.2; opacity:0.62;">GPQA Diamond</div></td>
	<td align="center" style="background: linear-gradient(135deg, rgba(83,104,243,0.16) 0%, rgba(77,110,244,0.16) 100%); vertical-align:middle;"><b>86.9</b></td>
	<td align="center" style="vertical-align:middle;">86.4</td>
	<td align="center" style="vertical-align:middle;">82.5</td>
	<td align="center" style="vertical-align:middle;">81.7</td>
	<td align="center" style="vertical-align:middle;">84.3</td>
	</tr>
	<tr>
	<td align="center" colspan="6" style="background: linear-gradient(90deg, rgba(124,58,237,0.45) 0%, rgba(99,102,241,0.42) 50%, rgba(59,130,246,0.45) 100%); font-weight: bold; height:32px; padding-top:2px; padding-bottom:2px;"><i>Agentic</i></td>
	</tr>
	<tr>
	<td style="text-align:left; padding:9px 10px 9px 18px;"><div style="font-size:15px; font-weight:600; line-height:1.22;">Multi-tool orchestration</div><div style="margin-top:4px; font-size:11px; font-weight:400; line-height:1.2; opacity:0.62;">MCP-Atlas</div></td>
	<td align="center" style="background: linear-gradient(135deg, rgba(77,110,244,0.16) 0%, rgba(71,117,244,0.16) 100%); vertical-align:middle;"><b>56.7</b></td>
	<td align="center" style="vertical-align:middle;">54.2</td>
	<td align="center" style="vertical-align:middle;">49.4</td>
	<td align="center" style="vertical-align:middle;">46.8</td>
	<td align="center" style="vertical-align:middle;">55.0</td>
	</tr>
	<tr>
	<td style="text-align:left; padding:9px 10px 9px 18px;"><div style="font-size:15px; font-weight:600; line-height:1.22;">Web browsing and research</div><div style="margin-top:4px; font-size:11px; font-weight:400; line-height:1.2; opacity:0.62;">BrowseComp</div></td>
	<td align="center" style="background: linear-gradient(135deg, rgba(71,117,244,0.16) 0%, rgba(65,123,245,0.16) 100%); vertical-align:middle;"><b>57.4</b></td>
	<td align="center" style="vertical-align:middle;">56.4</td>
	<td align="center" style="vertical-align:middle;">44.8</td>
	<td align="center" style="vertical-align:middle;">41.5</td>
	<td align="center" style="vertical-align:middle;">—</td>
	</tr>
	<tr>
	<td style="text-align:left; padding:9px 10px 9px 18px;"><div style="font-size:15px; font-weight:600; line-height:1.22;">Real-user agentic coding</div><div style="margin-top:4px; font-size:11px; font-weight:400; line-height:1.2; opacity:0.62;">ClawEval</div></td>
	<td align="center" style="background: linear-gradient(135deg, rgba(65,123,245,0.16) 0%, rgba(59,130,246,0.16) 100%); vertical-align:middle;"><b>67.8</b></td>
	<td align="center" style="vertical-align:middle;">66.5</td>
	<td align="center" style="vertical-align:middle;">63.1</td>
	<td align="center" style="vertical-align:middle;">53.2</td>
	<td align="center" style="vertical-align:middle;">48.5</td>
	</tr>
	</table>

	<p style="margin-top:12px;font-size:10px;opacity:0.7">
	* All results reported for OxCoder-9B are averaged over five independent runs. A dash (—) means the score was not reported for that model.<br/>
	* Terminal-Bench 2.1 (Terminus-2): We evaluate Terminal-Bench 2.1 using the Harbor/Terminus-2 framework with parser=json, temperature=1.0, top_p=1.0, and a 256K context window. Each run uses a 2-hour timeout with 32 CPU cores and 32GB RAM.<br/>
	* Terminal-Bench 2.1 (Claude Code): We evaluate Terminal-Bench 2.1 using Claude Code 2.1.126 with parser=json, temperature=1.0, top_p=1.0.<br/>
	* SWE-bench Verified and Pro: using the OpenHands harness with temp=1.0, top_p=0.95, 256K context window. Anti-hacking safeguards are applied throughout evaluation: Git history is removed from the local repository image to prevent access to prior solutions or commits; network access is disabled, preventing the model from retrieving external information or resources.<br/>
	* NL2Repo: with temperature=1.0, top_p=1.0, 256K context, 48K output. Access to the specified GitHub repositories and pip packages is blocked to prevent reward hacking.<br/>
	* HLE: Evaluated using GLM-5.3 as the judge model.<br/>
	* MCP-Atlas: Evaluated using GLM-5.3 as the judge model.<br/>
	* ClawEval: temp=0.6 and 256K context.<br/>
	* Baseline scores for Ornith-1.5-9B, Ornith-1.0-9B, Qwen3.5-9B and Gemma-4-31B are taken from the Ornith-1.5-9B model card and were produced under that team's evaluation settings, which differ from ours on some benchmarks (context window, timeouts and judge model).<br/>
	</p>

	---

	<div align="center">

	OxCoder-9B is developed by [OrionLLM](https://huggingface.co/OrionLLM) and released under the Apache 2.0 License.

	</div>