AINewsnow

B200 vs H100: how much faster is Blackwell for training?

In MLPerf Training v6.0, 64 NVIDIA B200 GPUs trained two LLM benchmarks 2.04 to 2.08 times faster than 64 NVIDIA H100s, a little under the 2.27x gap in their dense FP8 specs. Nodus lists the B200 at $5.50/hr and the H100 SXM at $2.60/hr, so a B200 must finish 2.115 times faster to cost less per run…

Read the full story at DEV Community — Machine Learning ↗

Timeline · 1 report

  1. 2026-11-10 21:39 · DEV Community — Machine Learning
    B200 vs H100: how much faster is Blackwell for training?

More stories

  1. Anthropic launches IPO roadshow targeting up to $1.4 trillion valuation — CNBC Technology
  2. Nvidia CEO Jensen Huang's affair with alleged Chinese spy draws national security scrutiny — New York Times AI
  3. Nvidia board probes CEO Jensen Huang over affair with alleged Chinese spy — South China Morning Post Tech
  4. Google adds TPU 8tx and 8i Max to eighth-generation AI chip lineup — Google: The Keyword
  5. AI training startup Mercor raises $750 million at $20 billion valuation — Economic Times Tech
  6. Hot inflation data revives December rate-hike expectations, weighing on stocks — Reuters AI
  7. Moonshot AI rented Nvidia B300 servers in China, filings show — Bloomberg AI
  8. Nvidia's General Counsel Intervened in US Probe of Megaspeed AI Chip Smuggling — Bloomberg AI

Get the daily brief of stories like this at 6:30 every morning →