DeepSeek-R1-Distill-Qwen-32B
‹ Back to models page
DeepSeek-R1-Distill-Qwen-32B is a distilled version of DeepSeek's R1 model, fine-tuned from the Qwen-2.5-32B base model. This model leverages knowledge distillation to retain robust reasoning capabilities. Delivering exceptional performance on mathematical and logical reasoning tasks, it achieves near-o1 level capabilities with faster response times. With its massive 128K context window, native tool use, and JSON mode support, it excels at complex problem-solving while maintaining the reasoning depth of much larger models.
Key Technical Specifications
Model Architecture
Built upon the Qwen-2.5-32B framework, the model is comprised of 32 billion parameters. The distillation process fine-tunes the base model using outputs from DeepSeek-R1, effectively transferring reasoning patterns.
Performance Metrics
The model demonstrates strong performance across various benchmarks:
- AIME 2024: Pass@1 score of 72.6
- MATH-500: Pass@1 score of 94.3
- CodeForces Rating: Achieved a rating of 1,691
Technical Details
| FEATURE | VALUE |
|---|---|
| Context Window (Tokens) | 128K |
| Max Output Tokens | 16,384 |
| Max File Size | - |
| Token Generation Speed | ~140 tps |
| Input Token Price | $0.69/1M tokens |
| Output Token Price | $0.69/1M tokens |
| Tool Use |  |
| JSON Mode | |
| Image Support |  |
Use Cases
Mathematical Problem-Solving
Effectively addresses complex mathematical queries, making it valuable for educational tools and research applications.
Coding Assistance
Supports code generation and debugging, beneficial for software development.
Logical Reasoning
Performs tasks requiring structured thinking and deduction, applicable in data analysis and strategic planning.
Best Practices
- Prompt Engineering: Set the temperature parameter between 0.5 and 0.7 (ideally 0.6) to prevent repetitive or incoherent outputs.
- System Prompt: Avoid adding a system prompt and include all instructions within the user prompt.
Get Started with DeepSeek-R1-Distill-Qwen-32B
Experience the reasoning capabilities of deepseek-r1-distill-qwen-32b with Groq speed now:
pip install groq1from groq import Groq
2client = Groq()
3completion = client.chat.completions.create(
4 model="deepseek-r1-distill-qwen-32b",
5 messages=[
6 {
7 "role": "user",
8 "content": "Explain why fast inference is critical for reasoning models"
9 }
10 ]
11)
12print(completion.choices[0].message.content)