Introduced DeepSeek-R1 optimizations for Blackwell, delivering 25x more revenue at 20x lower cost per token, compared with NVIDIA H100 just four weeks ago

Tuesday, February 25, 2025 1:22:50 AMEST
- Fueled by TensorRT DeepSeek optimizations for our Blackwell architecture, including FP4 performance with state-of-the-art production accuracy, it scored 99.8% of FP8 on MMLU general intelligence benchmark.

**Link: https://x.com/NVIDIAAIDev/status/1894172956726890623
~