From NVIDIA: 2025-01-31 12:00:01

The new DeepSeek-R1 model family is making waves in the AI community, offering advanced reasoning capabilities for problem-solving, math, and code tasks on local PCs. With up to 3,352 trillion operations per second, NVIDIA GeForce RTX 50 Series GPUs can run these models faster than any other PC hardware. These reasoning models focus on deep thought and reflection to solve complex problems and provide better results through test-time scaling. The DeepSeek-R1 family includes smaller student models distilled from a large 671-billion-parameter model, offering powerful reasoning capabilities locally on RTX AI PCs. GeForce RTX GPUs ensure peak performance for inference tasks, accelerating DeepSeek models on PCs. NVIDIA’s RTX AI platform provides access to DeepSeek-R1 capabilities on over 100 million RTX AI PCs worldwide, offering high performance, low latency, and increased privacy for AI tasks. Experience the power of DeepSeek-R1 and RTX AI PCs through a range of software tools for inference tasks and model fine-tuning.



Read more at NVIDIA: Accelerate DeepSeek Models With GeForce RTX 50 Series AI PCs