Instructions to use domyn/Domyn-Edge-Teacher-Math with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use domyn/Domyn-Edge-Teacher-Math with Transformers:

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

pipe = pipeline("text-generation", model="domyn/Domyn-Edge-Teacher-Math", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("domyn/Domyn-Edge-Teacher-Math", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use domyn/Domyn-Edge-Teacher-Math with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/domyn/Domyn-Edge-Teacher-Math

SGLang

How to use domyn/Domyn-Edge-Teacher-Math 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 "domyn/Domyn-Edge-Teacher-Math" \
    --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": "domyn/Domyn-Edge-Teacher-Math",
		"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 "domyn/Domyn-Edge-Teacher-Math" \
        --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": "domyn/Domyn-Edge-Teacher-Math",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use domyn/Domyn-Edge-Teacher-Math with Docker Model Runner:
```
docker model run hf.co/domyn/Domyn-Edge-Teacher-Math
```

Domyn Edge Teacher Math

Phased release. Domyn Edge is a rolling release in three bundles. This teacher is part of the second bundle, together with the other domain teachers and their post-training data. See Open resources for everything released so far.

This repository holds the math teacher of Domyn Edge: Domyn Edge SFT after GRPO on math problems with verifiable final answers. It is one of the domain teachers that score the student's rollouts token by token in the multi-teacher on-policy distillation (MOPD) stage of our post-training protocol. It is released for research and reproducibility; it is not a general-purpose assistant.

Model summary
How to use
Evaluation
Limitations
Open resources
License
Citation

Model summary

Domyn Edge Teacher Math is a 4B-parameter dense transformer with the same architecture and tokenizer as Domyn Edge SFT. It was trained in one stage on top of Domyn Edge SFT:

GRPO: asynchronous group relative policy optimization on math problems with a single verifiable final answer. The reward checks the final answer with math-verify: 1 for a correct answer, 0.5 for a correct answer not given in the format the prompt asks for, 0 otherwise.

The weights are stored in bfloat16.

How to use

The model loads and serves exactly like Domyn Edge SFT; see the Domyn Edge SFT card for the recommended environment, Transformers and vLLM instructions.

Chat template. The repository ships the Domyn Edge SFT chat template unchanged; it is the template used during GRPO training and in the evaluation below. It opens the system turn with a fixed identity line ("You are Domyn Edge, a model trained by Domyn.").

Output format. The model reasons first and ends with the final answer, usually in \boxed{}.

Evaluation

Teacher vs student on the general evaluation suite

Scores of this teacher next to those of the student, Domyn Edge SFT, with one benchmark per area, to check what the teacher keeps outside its training domain. All scores are percentages.

Group	Benchmark	Metric	Domyn Edge SFT	Teacher Math	Δ
Math	AIME 2026	avg@10	61.3	79.0	+17.7
Multilingual math	MGSM (de, en, es, fr, ru)	accuracy	85.8	87.4	+1.6
Knowledge and reasoning	GPQA Diamond	pass@1	56.3	59.6	+3.3
Instruction following	IFBench	prompt strict pass@1	24.1	25.0	+0.9
Function calling	BFCL v4, without web search (overall)	pass@1	36.5	36.2	-0.3
	BFCL v4 (non-live)	pass@1	51.9	53.3	+1.4
	BFCL v4 (live)	pass@1	75.9	75.1	-0.8
	BFCL v4 (multi-turn)	pass@1	35.9	38.8	+2.9
	BFCL v4 (memory)	pass@1	27.3	21.5	-5.8
Agentic	SWE-bench Verified	pass@1	33.6	29.6	-4.0
Code	LiveCodeBench (full set)*	pass@1	73.0	72.0	-1.0

*LiveCodeBench (full set) is all 1,055 problems of LiveCodeBench v6, released between May 2023 and April 2025.

Evaluation settings

Served with vLLM in bfloat16, with the chat template shipped in this repository.
Thinking on. Sampling: temperature 1.0, top-p 0.95, top-k 20, min-p 0, no presence or repetition penalty.

Math evaluation suite

Benchmark	Metric	Teacher Math	Qwen3.5 4B	Δ
AIME 2026	avg@10	79.0	82.9	-3.9
AIME 2025	avg@10	75.9	78.3	-2.4
HMMT Feb 2026	avg@10	65.9	63.3	+2.6
HMMT Feb 2025	avg@10	76.1	74.6	+1.5
GSM8K	accuracy	93.8	95.0	-1.2
MGSM (de, en, es, fr, ru)	accuracy	87.4	89.8	-2.4

Evaluation settings

Served with vLLM in bfloat16 at 128k context, thinking on.
Sampling: temperature 1.0, top-p 0.95, top-k 20, min-p 0, no presence, frequency or repetition penalty.
AIME and HMMT: 10 samples per problem, accuracy averaged over the samples (avg@10).

Limitations

A teacher, not an assistant. This model is one of the domain teachers of multi-teacher on-policy distillation (MOPD), the post-training stage in which the Domyn Edge student generates its own responses and each teacher scores them token by token in its own domain. Only math was trained beyond Domyn Edge SFT.

Long generations. On hard problems the model can reason for tens of thousands of tokens. Training capped responses at 64k tokens; cap the generation length.

Safety work is incomplete. No dedicated safety tuning or red teaming was done on this model.

Open resources

This teacher: domyn/Domyn-Edge-Teacher-Math
Its data: domyn/Domyn-Edge-Teacher-Math-Mix
Student SFT model: domyn/Domyn-Edge-SFT, and its mix domyn/Domyn-Edge-SFT-Mix
Collection: Domyn Edge on Hugging Face

License

This model is released under Apache 2.0.

Citation

@misc{domynedge2026,
  title  = {Domyn Edge},
  author = {Domyn},
  year   = {2026},
  note   = {Technical report, forthcoming in November 2026.},
}

Downloads last month: 4

Safetensors

Model size

4B params

Tensor type

BF16

Model tree for domyn/Domyn-Edge-Teacher-Math

Base model

domyn/Domyn-Edge-Base

Finetuned

domyn/Domyn-Edge-SFT

Finetuned

(6)

this model

Dataset used to train domyn/Domyn-Edge-Teacher-Math

Collection including domyn/Domyn-Edge-Teacher-Math

Domyn Edge

Collection

17 items • Updated 3 days ago

domyn
/

Domyn-Edge-Teacher-Math