Instructions to use zz1358m/Qwen3-30B-A3B-Agentic-ESOpt-Math with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use zz1358m/Qwen3-30B-A3B-Agentic-ESOpt-Math with Transformers:

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

pipe = pipeline("text-generation", model="zz1358m/Qwen3-30B-A3B-Agentic-ESOpt-Math")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("zz1358m/Qwen3-30B-A3B-Agentic-ESOpt-Math")
model = AutoModelForCausalLM.from_pretrained("zz1358m/Qwen3-30B-A3B-Agentic-ESOpt-Math", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.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(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use zz1358m/Qwen3-30B-A3B-Agentic-ESOpt-Math with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "zz1358m/Qwen3-30B-A3B-Agentic-ESOpt-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": "zz1358m/Qwen3-30B-A3B-Agentic-ESOpt-Math",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/zz1358m/Qwen3-30B-A3B-Agentic-ESOpt-Math

SGLang

How to use zz1358m/Qwen3-30B-A3B-Agentic-ESOpt-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 "zz1358m/Qwen3-30B-A3B-Agentic-ESOpt-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": "zz1358m/Qwen3-30B-A3B-Agentic-ESOpt-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 "zz1358m/Qwen3-30B-A3B-Agentic-ESOpt-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": "zz1358m/Qwen3-30B-A3B-Agentic-ESOpt-Math",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use zz1358m/Qwen3-30B-A3B-Agentic-ESOpt-Math with Docker Model Runner:
```
docker model run hf.co/zz1358m/Qwen3-30B-A3B-Agentic-ESOpt-Math
```

Qwen3-30B-A3B Agentic-ESOpt Math (theta20)

This is the full-parameter BF16 checkpoint obtained by deterministically replaying the first 20 Agentic-ESOpt updates (generations 0–19) on Qwen3-30B-A3B.

Reported result

Benchmark	Mean4	Pass4
DAPO	50.0	74.0
AIME2026	28.3	46.7

The source evaluation generated 16 samples per problem. The reported Mean4/Pass4 row is the group with sample_index 12–15. Mean4 is the mean score over those four samples; Pass4 is the mean per-problem maximum over those four samples. Exact unrounded values and the source evaluation records are included under repro/.

Replay configuration

Base model: Qwen/Qwen3-30B-A3B
Parameter scope: full (30,532,122,624 floating-point parameters)
Replayed updates: 20 (generations 0–19)
Population: 16
Alpha: 0.0005
Reward normalization: z-score (ddof=0, eps=1e-8)
Training sigma schedule: cosine, 0.001 to 0.0005
Export dtype: BF16

es_replay_export_manifest.json records every replayed generation and verifies that the recomputed normalization weights exactly equal the weights stored in the history. repro/es_history_theta20.json is the exact history prefix used to construct this checkpoint.