Workstation

• Large language model platform
• Dual Gigabit Ethernet
• 8-slot graphics card support

With 2 RTX-5090 graphics cards to run the deepseek-r1:70b AI model generates about 30 tokens per second.