Infrastructure & DevOps

Compute

Last updated 2026-09-01 · benchmark measured 2026-09-01 — deterministic & reproducible

Low-cost NVIDIA H100 and H200 GPU cloud for AI training and inference.

Is Compute production-ready?

Legit.Show scores Compute 72 out of 100 — the simple average of its 7 measured frames. Legit.Show ran its deterministic 7-Frame production-readiness benchmark on Compute (public-surface assessment), measured from the public surface with no LLM in the scoring path. Its strongest frame is Reliability; its weakest is Privacy. Every frame it averages is published with its evidence on the Legit.Show listing.

The 7 Frames

• Performance — 65/100
• Accessibility — 96/100
• Security — 80/100
• Privacy — 0/100
• Reliability — 100/100
• Standards — 92/100
• Discoverability — 70/100

What we measured

• Security headers present: CSP, X-Frame-Options, X-Content-Type-Options, Referrer-Policy.
• No HSTS.
• Served over HTTPS with a valid certificate.
• Real Lighthouse performance run — 603 ms to first byte.
• Returns a proper 404 for unknown routes.
• 3 of 3 sampled routes reachable.
• No privacy policy found.
• Sets cookies / loads scripts with no consent prompt.

Who it's for

AI researchers · ML engineers · Teams training models · Inference workloads · Cost-conscious compute users

Pricing

H100 on-demand: $1.931/GPU-hour, H100 spot: $1.246/GPU-hour, H200 on-demand: $2.917/GPU-hour, H200 spot: $1.855/GPU-hour

Visit Compute → · Alternatives to Compute → · How this was measured →