---
title: Qwen3-32B-AWQ
canonical_url: "https://www.modelscope.cn/models/swift/Qwen3-32B-AWQ"
md_url: "https://www.modelscope.cn/models/swift/Qwen3-32B-AWQ.md"
repository: swift/Qwen3-32B-AWQ
last_updated: 2025-05-14
license: "Apache License 2.0"
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - qwen3
architectures:
  - Qwen3ForCausalLM
base_model:
  - Qwen/Qwen3-32B
base_model_relation: quantized
parameters: 5.7B
tensor_type:
  - I32
  - F16
library_name:
  - pytorch
  - safetensors
frameworks:
  - Pytorch
inference_backends:
  - "deploy_task text/emb"
  - "lmdeploy_turbomind 0.9.1"
  - "sglang 0.5.2"
  - "vllm 0.9.2"
downloads: 6933
stars: 6
---

# Qwen3-32B-AWQ

> Qwen3-32B-AWQ - swift 在 ModelScope 开源的模型。The model has undergone AWQ int4 quantization using the ms-swift framework.

swift/Qwen3-32B-AWQ 是 ModelScope 魔搭社区上的 5.7B 参数text-generation模型，采用 Apache License 2.0 许可，基于 Qwen/Qwen3-32B 构建，可用 deploy_task text/emb、lmdeploy_turbomind 0.9.1、sglang 0.5.2 部署。

- **Repository**: swift/Qwen3-32B-AWQ
- **License**: Apache License 2.0
- **Tasks**: text-generation
- **Parameters**: 5.7B
- **Base model**: Qwen/Qwen3-32B
- **Inference backends**: deploy_task text/emb, lmdeploy_turbomind 0.9.1, sglang 0.5.2, vllm 0.9.2
- **Downloads**: 6933
- **Stars**: 6
- **Last updated**: 2025-05-14

Source: https://www.modelscope.cn/models/swift/Qwen3-32B-AWQ

---

## Inference
```python
import torch
from modelscope import AutoModelForCausalLM, AutoTokenizer

model_name = "swift/Qwen3-32B-AWQ"

# load the tokenizer and the model
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype=torch.float16,
    device_map="auto"
)

# prepare the model input
prompt = "Give me a short introduction to large language model."
messages = [
    {"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True,
    enable_thinking=True # Switches between thinking and non-thinking modes. Default is True.
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)

# conduct text completion
generated_ids = model.generate(
    **model_inputs,
    max_new_tokens=32768
)
output_ids = generated_ids[0][len(model_inputs.input_ids[0]):].tolist() 

# parsing thinking content
try:
    # rindex finding 151668 (</think>)
    index = len(output_ids) - output_ids[::-1].index(151668)
except ValueError:
    index = 0

thinking_content = tokenizer.decode(output_ids[:index], skip_special_tokens=True).strip("\n")
content = tokenizer.decode(output_ids[index:], skip_special_tokens=True).strip("\n")

print("thinking content:", thinking_content)
print("content:", content)
```

## Quantization

The model has undergone AWQ int4 quantization using the [ms-swift](https://github.com/modelscope/ms-swift) framework.

If you have fine-tuned the model and wish to quantize the fine-tuned version, you can refer to the following quantization scripts:

- Dense Model Quantization Script: [View Here](https://github.com/modelscope/ms-swift/blob/main/examples/export/quantize/awq.sh)
- MoE Model Quantization Script: [View Here](https://github.com/modelscope/ms-swift/blob/main/examples/export/quantize/moe/awq.sh)

With these scripts, you can easily complete the quantization process for the model.


## Evaluation

We evaluate the quality of this AWQ quantization with [EvalScope](https://github.com/modelscope/evalscope). For the best practice for evaluating Qwen3 models, one may refer to the following:
- [最佳实践](https://evalscope.readthedocs.io/zh-cn/latest/best_practice/qwen3.html)
- [Best Practice](https://evalscope.readthedocs.io/en/latest/best_practice/qwen3.html)

Performance of Qwen3-32B-AWQ is evaluated on our mixed-benchmark of [Qwen3 Evaluation Collection](https://modelscope.cn/datasets/modelscope/EvalScope-Qwen3-Test), with the results listed below:


> The performance comparison of Qwen3-32B-AWQ and Qwen3-32B

| task_type   | dataset_name    | metric                  | average_score(AWQ) | average_score(without AWQ) | count |
|-------------|-----------------|-------------------------|--------------------|----------------------------|-------|
| exam        | MMLU-Pro        | AverageAccuracy         | 0.7906             | 0.8018                     | 12032 |
| exam        | MMLU-Redux      | AverageAccuracy         | 0.8918             | 0.893                      | 5700  |
| exam        | C-Eval          | AverageAccuracy         | 0.8834             | 0.8915                     | 1346  |
| instruction | IFEval          | inst_level_strict_acc   | 0.8842             | 0.8802                     | 541   |
| instruction | IFEval          | inst_level_loose_acc    | 0.915              | 0.9125                     | 541   |
| instruction | IFEval          | prompt_level_loose_acc  | 0.8725             | 0.8595                     | 541   |
| instruction | IFEval          | prompt_level_strict_acc | 0.8336             | 0.8189                     | 541   |
| math        | MATH-500        | AveragePass@1           | 0.932              | 0.942                      | 500   |
| knowledge   | GPQA            | AveragePass@1           | 0.6566             | 0.6465                     | 198   |
| code        | LiveCodeBench   | Pass@1                  | 0.5                | 0.544                      | 182   |
| exam        | iQuiz           | AverageAccuracy         | 0.8                | 0.775                      | 120   |
| math        | AIME 2024       | AveragePass@1           | 0.7333             | 0.7667                     | 30    |
| math        | AIME 2025       | AveragePass@1           | 0.5667             | 0.6667                     | 30    |

### Conclusion

As we can see from the comparison above, evaluatoin results across different tasks and datasets suggest that our quantized-version with AWQ exihibit minimum fluctuation on model performance. In fact, for most benchmarks, AWQ version performs mostly on-par with the original version, except for math-related benchmarks (such as AIME2024/AIME2025) where performance degration is a bit more noticable.

Please also note that the result for the Qwen3-32B model is **not** Qwen offical. It is done in the same setup that we used to evaluate Qwen3-32B-AWQ. You may reproduce this evaluation result by following our best-practice guides above.
