AI & Agents

Bench-bench

Last updated 2026-08-18 · benchmark measured 2026-08-18 — deterministic & reproducible

Bench-bench benchmarks AI models by simulating a weightlifting training program.

Is Bench-bench production-ready?

Legit.Show scores Bench-bench 58 out of 100 — the simple average of its 7 measured frames. Legit.Show ran its deterministic 7-Frame production-readiness benchmark on Bench-bench (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 — 55/100
• Accessibility — 90/100
• Security — 25/100
• Privacy — 25/100
• Reliability — 92/100
• Standards — 67/100
• Discoverability — 50/100

What we measured

• No Content-Security-Policy and no HSTS.
• Served over HTTPS with a valid certificate.
• Real Lighthouse performance run — 105 ms to first byte.
• Returns a proper 404 for unknown routes.
• 0 of 0 sampled routes reachable.
• No privacy policy found.
• Sets cookies / loads scripts with no consent prompt.
• Discoverable: canonical URL.

Who it's for

AI researchers · Model developers · AI enthusiasts · Fitness enthusiasts · Benchmark researchers

Pricing

$115 for v0.2 public run (4 models, 40 simulated years)

Visit Bench-bench → · Alternatives to Bench-bench → · How this was measured →