DeepSeek R1 Distill Qwen 1.5B
Zero-eval
by DeepSeek
+
+
+
+
About
DeepSeek-R1-Distill-Qwen-1.5B was created through distillation into an ultra-compact Qwen architecture, designed to enable reasoning capabilities on resource-constrained devices. Built with just 1.5 billion parameters, it brings advanced analytical techniques to edge computing and mobile scenarios.
+
+
+
+
Timeline
AnnouncedJan 20, 2025
ReleasedJan 20, 2025
+
+
+
+
Specifications
Training Tokens14.8T
+
+
+
+
License & Family
License
MIT
Performance Overview
Performance metrics and category breakdown
Overall Performance
4 benchmarks
Average Score
46.8%
Best Score
83.9%
High Performers (80%+)
1+
+
+
+
All Benchmark Results for DeepSeek R1 Distill Qwen 1.5B
Complete list of benchmark scores with detailed information
| MATH-500 | text | 0.84 | 83.9% | Self-reported | |
| AIME 2024 | text | 0.53 | 52.7% | Self-reported | |
| GPQA | text | 0.34 | 33.8% | Self-reported | |
| LiveCodeBench | text | 0.17 | 16.9% | Self-reported |
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+