Nvidia Groq 3 LPX Hits 3,400 Tokens/Sec
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Nvidia says its Groq 3 LPX racks delivered 3,400 tokens per second in an Artificial Analysis benchmark running Gemma 4 31B with a 100,000-token input sequence
- Nvidia disclosed the performance figures on Monday, framing the results as validation of its $20 billion bet on Groq's LPU technology
Why it matters: Nvidia is publicly backing its $20 billion Groq LPU bet with hard throughput numbers: 3,400 tokens/sec on a 100,000-token Gemma 4 31B input. The Monday disclosure gives Nvidia its first concrete performance data point for the Groq-derived LPU stack, moving the technology from acquisition narrative to measurable benchmark territory.
Ask SkimNews
