Instructions to use Alibaba-Apsara/DASD-30B-A3B-Thinking-Preview with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Alibaba-Apsara/DASD-30B-A3B-Thinking-Preview with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="Alibaba-Apsara/DASD-30B-A3B-Thinking-Preview") messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("Alibaba-Apsara/DASD-30B-A3B-Thinking-Preview") model = AutoModelForCausalLM.from_pretrained("Alibaba-Apsara/DASD-30B-A3B-Thinking-Preview") 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 Alibaba-Apsara/DASD-30B-A3B-Thinking-Preview with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "Alibaba-Apsara/DASD-30B-A3B-Thinking-Preview" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Alibaba-Apsara/DASD-30B-A3B-Thinking-Preview", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/Alibaba-Apsara/DASD-30B-A3B-Thinking-Preview
- SGLang
How to use Alibaba-Apsara/DASD-30B-A3B-Thinking-Preview 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 "Alibaba-Apsara/DASD-30B-A3B-Thinking-Preview" \ --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": "Alibaba-Apsara/DASD-30B-A3B-Thinking-Preview", "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 "Alibaba-Apsara/DASD-30B-A3B-Thinking-Preview" \ --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": "Alibaba-Apsara/DASD-30B-A3B-Thinking-Preview", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use Alibaba-Apsara/DASD-30B-A3B-Thinking-Preview with Docker Model Runner:
docker model run hf.co/Alibaba-Apsara/DASD-30B-A3B-Thinking-Preview
Update README.md
Browse filesupdate the url of the technical report
README.md
CHANGED
|
@@ -12,24 +12,19 @@ datasets:
|
|
| 12 |
|
| 13 |
|
| 14 |
[](https://github.com/D2I-ai/dasd-thinking) 
|
| 15 |
-
|
| 16 |
|
| 17 |
|
| 18 |
|
| 19 |
[](https://huggingface.co/Alibaba-Apsara/DASD-4B-Thinking) 
|
| 20 |
-
[](https://www.modelscope.cn/models/Alibaba-Apsara/DASD-4B-Thinking) 
|
| 21 |
|
| 22 |
|
| 23 |
[](https://huggingface.co/Alibaba-Apsara/DASD-30B-A3B-Thinking-Preview) 
|
| 24 |
-
[](https://www.modelscope.cn/models/Alibaba-Apsara/DASD-30B-A3B-Thinking-Preview) 
|
| 25 |
-
|
| 26 |
|
| 27 |
|
| 28 |
[](https://huggingface.co/datasets/Alibaba-Apsara/Superior-Reasoning-SFT-gpt-oss-120b) 
|
| 29 |
-
[](https://www.modelscope.cn/datasets/Alibaba-Apsara/Superior-Reasoning-SFT-gpt-oss-120b) 
|
| 30 |
|
| 31 |
[](https://huggingface.co/datasets/Alibaba-Apsara/Superior-Reasoning-SFT-gpt-oss-120b-Logprob) 
|
| 32 |
-
[](https://www.modelscope.cn/datasets/Alibaba-Apsara/Superior-Reasoning-SFT-gpt-oss-120b-Logprob) 
|
| 33 |
|
| 34 |
|
| 35 |
| Model | AIME25 | LiveCodeBench v6 | GPQA-D | Average |
|
|
@@ -46,7 +41,7 @@ We release DASD-30B-A3B-Thinking-Preview, a highly capable 30B Mixture-of-Expert
|
|
| 46 |
|
| 47 |
> Note1: To demonstrate the scalability and efficiency of our data recipe, this preview model was trained only on the first-stage (Low-Temperature) dataset (~105K samples) derived from our 4B pipeline, without any re-curation or additional RL. Even with this lightweight recipe, it achieves excellent performance among open MoE models.
|
| 48 |
|
| 49 |
-
> Note2: This model (DASD-30B-A3B-Thinking-Preview) is a preliminary research artifact trained only on the first stage (Low-Temperature Sampling) of our pipeline to demonstrate the scalability of our data recipe. For the fully trained model and complete methodology, please refer to [DASD-4B-Thinking](https://huggingface.co/Alibaba-Apsara/DASD-4B-Thinking) and our [Technical Report](https://
|
| 50 |
|
| 51 |
|
| 52 |
|
|
@@ -136,12 +131,13 @@ While DASD-30B-A3B-Thinking-Preview demonstrates remarkable performance across m
|
|
| 136 |
DASD-Thinking is developed by Alibaba Cloud, as part of our mission to advance open, efficient, and trustworthy reasoning systems. If you find this work useful in your research or applications, please cite our technical report.
|
| 137 |
|
| 138 |
```bibtex
|
| 139 |
-
@
|
| 140 |
title={Distribution-Aligned Sequence Distillation for Superior Long-CoT Reasoning},
|
| 141 |
author={Yan, Shaotian and Liu, Kaiyuan and Shen, Chen and Wang, Bing and Fan, Sinan and Zhang, Jun and Wu, Yue and Wang, Zheng and Ye, Jieping},
|
| 142 |
year={2026},
|
| 143 |
-
|
| 144 |
-
}
|
|
|
|
| 145 |
|
| 146 |
@article{liu2025where,
|
| 147 |
title={Where Did This Sentence Come From? Tracing Provenance in LLM Reasoning Distillation},
|
|
|
|
| 12 |
|
| 13 |
|
| 14 |
[](https://github.com/D2I-ai/dasd-thinking) 
|
| 15 |
+
<a href="https://arxiv.org/abs/2601.09088" target="_blank"><img src="https://img.shields.io/badge/Technical Report-b5212f.svg?logo=arxiv" height="21px"></a>
|
| 16 |
|
| 17 |
|
| 18 |
|
| 19 |
[](https://huggingface.co/Alibaba-Apsara/DASD-4B-Thinking) 
|
|
|
|
| 20 |
|
| 21 |
|
| 22 |
[](https://huggingface.co/Alibaba-Apsara/DASD-30B-A3B-Thinking-Preview) 
|
|
|
|
|
|
|
| 23 |
|
| 24 |
|
| 25 |
[](https://huggingface.co/datasets/Alibaba-Apsara/Superior-Reasoning-SFT-gpt-oss-120b) 
|
|
|
|
| 26 |
|
| 27 |
[](https://huggingface.co/datasets/Alibaba-Apsara/Superior-Reasoning-SFT-gpt-oss-120b-Logprob) 
|
|
|
|
| 28 |
|
| 29 |
|
| 30 |
| Model | AIME25 | LiveCodeBench v6 | GPQA-D | Average |
|
|
|
|
| 41 |
|
| 42 |
> Note1: To demonstrate the scalability and efficiency of our data recipe, this preview model was trained only on the first-stage (Low-Temperature) dataset (~105K samples) derived from our 4B pipeline, without any re-curation or additional RL. Even with this lightweight recipe, it achieves excellent performance among open MoE models.
|
| 43 |
|
| 44 |
+
> Note2: This model (DASD-30B-A3B-Thinking-Preview) is a preliminary research artifact trained only on the first stage (Low-Temperature Sampling) of our pipeline to demonstrate the scalability of our data recipe. For the fully trained model and complete methodology, please refer to [DASD-4B-Thinking](https://huggingface.co/Alibaba-Apsara/DASD-4B-Thinking) and our [Technical Report](https://arxiv.org/abs/2601.09088).
|
| 45 |
|
| 46 |
|
| 47 |
|
|
|
|
| 131 |
DASD-Thinking is developed by Alibaba Cloud, as part of our mission to advance open, efficient, and trustworthy reasoning systems. If you find this work useful in your research or applications, please cite our technical report.
|
| 132 |
|
| 133 |
```bibtex
|
| 134 |
+
@article{yan2026dasd,
|
| 135 |
title={Distribution-Aligned Sequence Distillation for Superior Long-CoT Reasoning},
|
| 136 |
author={Yan, Shaotian and Liu, Kaiyuan and Shen, Chen and Wang, Bing and Fan, Sinan and Zhang, Jun and Wu, Yue and Wang, Zheng and Ye, Jieping},
|
| 137 |
year={2026},
|
| 138 |
+
journal={arXiv preprint arXiv:2601.09088},
|
| 139 |
+
url={https://arxiv.org/abs/2601.09088}
|
| 140 |
+
}
|
| 141 |
|
| 142 |
@article{liu2025where,
|
| 143 |
title={Where Did This Sentence Come From? Tracing Provenance in LLM Reasoning Distillation},
|