DeepSeek
DeepSeek V4.1 Flash Runs at 30–40 TPS on Four RTX 3090s, but Setup Is Still Demanding
A Reddit user ran DeepSeek V4.1 Flash on four RTX 3090s with Maya, hitting 30–40 tokens per second at 524K context.
1 article tagged with this topic
A Reddit user ran DeepSeek V4.1 Flash on four RTX 3090s with Maya, hitting 30–40 tokens per second at 524K context.