dusersad12 / OrionNet-Final-Checkpoint

OrionNet

---

1. Introduction

OrionNet is the latest release from Stellar Labs, built on a redesigned transformer backbone that scales efficiently with data and compute. This version introduces a refined mixture-of-experts routing strategy and an expanded context window, delivering stronger multi-step reasoning while keeping inference latency comparable to dense models of similar size. Extensive internal evaluations confirm that OrionNet surpasses its predecessor across mathematics, code, and general knowledge benchmarks.

Compared to the previous OrionNet release, the current checkpoint shows markedly improved long-context handling. On the MATH-500 suite, accuracy rose from 63% to 71%, and average response length grew from 9K to 18K tokens, reflecting deeper chain-of-thought before answering. Hallucination rates on the TruthfulQA probe dropped by 12 percentage points.

Beyond reasoning, this release also ships a more robust tool-use interface and lower latency for function-calling workloads.

2. Evaluation Results

Comprehensive Benchmark Results

Overall Performance Summary

OrionNet demonstrates competitive performance across all evaluated benchmark categories, with particularly strong results in instruction following and safety evaluation.

3. Chat Website & API Platform

We provide a chat interface and API for interacting with OrionNet. Please visit the Stellar Labs website for more information.

4. How to Run Locally

Please refer to our code repository for instructions on running OrionNet locally.

Key changes from the previous version:

1. System prompts are fully supported.
2. No special tokens are needed to activate the model's reasoning mode.

The architecture of OrionNet-Small matches the base model and shares the same tokenizer. It can be used as a drop-in replacement for the base model.

System Prompt

We recommend the following system prompt with a specific date.

You are OrionNet, a helpful AI assistant.
Today is {current date}.

For example,

You are OrionNet, a helpful AI assistant.
Today is March 15, 2026, Sunday.

Temperature

We recommend setting the temperature parameter $T_{model}$ to 0.7.

Prompts for File Uploading and Web Search

For file uploading, follow the template below where {file_name}, {file_content} and {question} are arguments.

file_template = \
"""[file name]: {file_name}
[file content begin]
{file_content}
[file content end]
{question}"""

For web search enhanced generation, use the following prompt template where {search_results}, {cur_date}, and {question} are arguments.

search_answer_en_template = \
'''# The following contents are the search results related to the user's message:
{search_results}
In the search results I provide to you, each result is formatted as [webpage X begin]...[webpage X end], where X represents the numerical index of each article. Please cite the context at the end of the relevant sentence when appropriate. Use the citation format [citation:X] in the corresponding part of your answer. If a sentence is derived from multiple contexts, list all relevant citation numbers, such as [citation:3][citation:5]. Be sure not to cluster all citations at the end; instead, include them in the corresponding parts of the answer.
When responding, please keep the following points in mind:
- Today is {cur_date}.
- Not all content in the search results is closely related to the user's question. You need to evaluate and filter the search results based on the question.
- For listing-type questions (e.g., listing all flight information), try to limit the answer to 10 key points and inform the user that they can refer to the search sources for complete information. Prioritize providing the most complete and relevant items in the list. Avoid mentioning content not provided in the search results unless necessary.
- For creative tasks (e.g., writing an essay), ensure that references are cited within the body of the text, such as [citation:3][citation:5], rather than only at the end of the text. You need to interpret and summarize the user's requirements, choose an appropriate format, fully utilize the search results, extract key information, and generate an answer that is insightful, creative, and professional. Extend the length of your response as much as possible, addressing each point in detail and from multiple perspectives, ensuring the content is rich and thorough.
- If the response is lengthy, structure it well and summarize it in paragraphs. If a point-by-point format is needed, try to limit it to 5 points and merge related content.
- For objective Q&A, if the answer is very brief, you may add one or two related sentences to enrich the content.
- Choose an appropriate and visually appealing format for your response based on the user's requirements and the content of the answer, ensuring strong readability.
- Your answer should synthesize information from multiple relevant webpages and avoid repeatedly citing the same webpage.
- Unless the user requests otherwise, your response should be in the same language as the user's question.
# The user's message is:
{question}'''

5. License

This code repository is licensed under the Apache 2.0 License . The use of OrionNet models is also subject to the Apache 2.0 License . The model series supports commercial use and distillation.

6. Contact

If you have any questions, please raise an issue on our GitHub repository or contact us at [email protected] .