Skip to main content
usedby
Search tools, companies

How Mistral AI uses CoreWeave

2.5x

“train models 2.5x faster than previously experienced on NVIDIA H200s”

As published on coreweave.com, October 2026. Captured by usedby on Oct 7, 2026.

What happened

Mistral AI trains and serves its large language models on CoreWeave GPU clusters (H100, H200 and GB200), running training and inference on the same clusters with Slurm on Kubernetes. It also relies on CoreWeave's observability tooling, automated health checks and 24/7 support.

Summary written by usedby from the source page, in English. The figures are those of CoreWeave and Mistral AI, not ours.

With the support of CoreWeave infrastructure, Mistral AI was able to train models 2.5x faster than previously experienced on NVIDIA H200s—and also vastly exceeded the performance, efficiency, and speed they experienced on NVIDIA H100s.

From the page. coreweave.com, captured Oct 7, 2026

What we don’t see with CoreWeave are all of the many interruptions or issues that we would see elsewhere.

Timothée Lacroix, CTO, Mistral AI. Source, captured Oct 7, 2026

[Our models] were trained 100% on CoreWeave infrastructure. I think not being on that kind of infrastructure would’ve delayed us by at least a few months.

Timothée Lacroix, CTO, Mistral AI. Source, captured Oct 7, 2026

What the story claims, and what we checked

We compared the story with its live page on Oct 7, 2026.

  • The figure: 2.5xCheckedPrinted word for word on the page, near the name of Mistral AI.
  • The passage quoted aboveCheckedCopied word for word from the page, near the name of Mistral AI.
  • The numbers in our summaryCheckedEach one is printed on the page.
  • The publication dateCheckedRead from the page’s own metadata, never guessed.
  • Mistral AI uses CoreWeaveCheckedConfirmed line. Latest check across sources: Oct 7, 2026.
  • The result itselfNot checkedWe quote it; we did not measure it.

Same company, same tool or same industry.

80%“Greater than 80% faster speed and enhanced performance results on NVIDIA GB200 NVL72”IBM uses CoreWeave. Another customer of CoreWeave1B+“1B+ Simultaneous viewers”Miris uses CoreWeave. Another customer of CoreWeave75%“saving 75% in cloud costs”NovelAI uses CoreWeave. Another customer of CoreWeave