Instructions to use dusersad12/MyNewModel-ReleaseRepo with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use dusersad12/MyNewModel-ReleaseRepo with Transformers:

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

pipe = pipeline("text-generation", model="dusersad12/MyNewModel-ReleaseRepo")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("dusersad12/MyNewModel-ReleaseRepo")
model = AutoModelForCausalLM.from_pretrained("dusersad12/MyNewModel-ReleaseRepo", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use dusersad12/MyNewModel-ReleaseRepo with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "dusersad12/MyNewModel-ReleaseRepo"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "dusersad12/MyNewModel-ReleaseRepo",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/dusersad12/MyNewModel-ReleaseRepo

SGLang

How to use dusersad12/MyNewModel-ReleaseRepo 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 "dusersad12/MyNewModel-ReleaseRepo" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "dusersad12/MyNewModel-ReleaseRepo",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "dusersad12/MyNewModel-ReleaseRepo" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "dusersad12/MyNewModel-ReleaseRepo",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use dusersad12/MyNewModel-ReleaseRepo with Docker Model Runner:
```
docker model run hf.co/dusersad12/MyNewModel-ReleaseRepo
```

MyNewModel

1. Introduction

MyNewModel is the newest release from our small team, trained with a longer post-training schedule and a revised data mixture than the release before it. The biggest gains this round are in multi-step reasoning and tool use: on competition-style math the model plans more carefully before answering, and on agentic benchmarks it recovers from failed tool calls instead of giving up. It stays competitive with considerably larger models on general knowledge tasks.

A concrete example of the reasoning jump: on a held-out set of olympiad-level algebra problems, the previous release solved 61% of problems, while MyNewModel solves 79%. Part of the gain comes from spending more thinking tokens per problem — the average went from roughly 9K to 21K tokens on the hardest tier.

Beyond raw reasoning, this release cuts the hallucination rate on factual short-form questions roughly in half and ships a much more reliable function-calling format.

2. Evaluation Results

Comprehensive Benchmark Results

	Benchmark	ModelA	ModelB	ModelA-v2	MyNewModel
Core Reasoning Tasks	Math Reasoning	0.482	0.501	0.517	0.687
	Logical Reasoning	0.755	0.772	0.788	0.879
	Common Sense	0.693	0.688	0.704	0.783
Language Understanding	Reading Comprehension	0.641	0.659	0.672	0.748
	Question Answering	0.553	0.571	0.584	0.619
	Text Classification	0.791	0.804	0.812	0.837
	Sentiment Analysis	0.748	0.762	0.779	0.818
Generation Tasks	Code Generation	0.592	0.608	0.621	0.717
	Creative Writing	0.561	0.549	0.583	0.658
	Dialogue Generation	0.601	0.617	0.628	0.686
	Summarization	0.718	0.731	0.742	0.792
Specialized Capabilities	Translation	0.759	0.776	0.781	0.815
	Knowledge Retrieval	0.634	0.649	0.661	0.688
	Instruction Following	0.712	0.728	0.736	0.750
	Safety Evaluation	0.695	0.678	0.702	0.773

Overall Performance Summary

Across the fifteen benchmark categories we track, MyNewModel posts its largest margins on the reasoning and code clusters, while holding steady on the language-understanding suite.

3. Chat Website & API Platform

We run a hosted chat playground and a metered API for MyNewModel. Sign-up details and rate limits are on our official website.

4. How to Run Locally

The training code, inference server, and quantized variants live in our code repository. Here we only note the usage points that changed with this release.

Compared with the previous release, keep the following in mind:

A system prompt is now recommended for best results.
Special control tokens that used to be required to enter thinking mode are no longer needed; the model decides on its own.

The model architecture of MyNewModel-Small matches its base model exactly, and it reuses the tokenizer of the main MyNewModel release, so existing serving stacks work unchanged.

System Prompt

Pair the model with a system prompt that includes the current date.

You are MyNewModel, an AI assistant that helps people.
Today is {current date}.

For example,

You are MyNewModel, an AI assistant that helps people.
Today is September 27, 2026, Sunday.

Temperature

We recommend setting the temperature parameter $T_{model}$ to 0.55.

Prompts for File Uploading and Web Search

For file uploading, please follow the template to create prompts, 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, we recommend 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 the answer 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 repository is released under the Apache 2.0 License. Use of the MyNewModel weights is also governed by the Apache 2.0 License. Commercial deployment and distillation are both allowed.

6. Contact

Questions or bug reports are welcome on our GitHub repository, or reach the team at hello@mynewmodel.ai.

Downloads last month: -