DeepSeek-R1
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
DeepSeek-R1 Experiment 1
100F1 Score (Ga)
40
DeepSeek-R1-Distill-Qwen-1.5B
89.8P_safeguarded (Safety-Quality Score)
17
DeepSeek-R1-Distill-Llama-8B 16K Generation Length
8Max Batch Size
10
DeepSeek-R1-Distill-Llama-8B 8K Generation Length
8Max Batch Size
10
DeepSeek-R1-Distill
1.997Score
6
DeepSeek-R1-Distill-Llama-8B serving environment
0.08TTFT (s)
6
DeepSeek-R1 Experiment 2
0.99BERT Score
5
DeepSeek-R1-Distill-LLaMA-8B 64K Context
24.1Latency (ms)
4