BREAKING
NVIDIA Pushes 'Cost Per Token' Metric
0$
per 1M tokens (Blackwell)
0$
per 1M tokens (Hopper)
0x
cost reduction
Tokens Per Second, Per GPU
Blackwell6000
Hopper90
0x
perf uplift in a month
0x
peak throughput gain
Partners Report Early Gains
Basetenreasoning
Serves DeepSeek V4 Pro on Blackwell
Up to 50% more tokens per second
Hippocratic AIscale
30% throughput lift
Sub-half-second response, 10M calls
Can Rivals Close the Gap
AI NEWS BLITZ
NVIDIA says the key AI metric is now cost per token, not raw chip specs.