<?xml version="1.0" encoding="UTF-8"?>
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
<url>
<loc>https://www.gradiahq.com</loc>
<changefreq>weekly</changefreq>
<priority>1</priority>
</url>
<url>
<loc>https://www.gradiahq.com/about</loc>
<changefreq>monthly</changefreq>
<priority>0.7</priority>
</url>
<url>
<loc>https://www.gradiahq.com/solutions</loc>
<changefreq>monthly</changefreq>
<priority>0.8</priority>
</url>
<url>
<loc>https://www.gradiahq.com/solutions/agent-evaluation</loc>
<changefreq>monthly</changefreq>
<priority>0.8</priority>
</url>
<url>
<loc>https://www.gradiahq.com/solutions/workflow-efficiency</loc>
<changefreq>monthly</changefreq>
<priority>0.8</priority>
</url>
<url>
<loc>https://www.gradiahq.com/solutions/benchmark-audit</loc>
<changefreq>monthly</changefreq>
<priority>0.8</priority>
</url>
<url>
<loc>https://www.gradiahq.com/governance</loc>
<changefreq>monthly</changefreq>
<priority>0.9</priority>
</url>
<url>
<loc>https://www.gradiahq.com/observer</loc>
<changefreq>monthly</changefreq>
<priority>0.7</priority>
</url>
<url>
<loc>https://www.gradiahq.com/observer/get-started</loc>
<changefreq>monthly</changefreq>
<priority>0.7</priority>
</url>
<url>
<loc>https://www.gradiahq.com/observer/downloads</loc>
<changefreq>monthly</changefreq>
<priority>0.7</priority>
</url>
<url>
<loc>https://www.gradiahq.com/research</loc>
<changefreq>monthly</changefreq>
<priority>0.8</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog</loc>
<changefreq>weekly</changefreq>
<priority>0.8</priority>
</url>
<url>
<loc>https://www.gradiahq.com/benchmarks</loc>
<changefreq>weekly</changefreq>
<priority>0.8</priority>
</url>
<url>
<loc>https://www.gradiahq.com/universes</loc>
<changefreq>weekly</changefreq>
<priority>0.8</priority>
</url>
<url>
<loc>https://www.gradiahq.com/explainers</loc>
<changefreq>monthly</changefreq>
<priority>0.8</priority>
</url>
<url>
<loc>https://www.gradiahq.com/explainers/how-gradia-works</loc>
<changefreq>monthly</changefreq>
<priority>0.8</priority>
</url>
<url>
<loc>https://www.gradiahq.com/explainers/gradia-technical</loc>
<changefreq>monthly</changefreq>
<priority>0.8</priority>
</url>
<url>
<loc>https://www.gradiahq.com/explainers/reward-hacking-rl-loop</loc>
<changefreq>monthly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/explainers/wind-tunnel</loc>
<changefreq>monthly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/explainers/conditionally-approved</loc>
<changefreq>monthly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/explainers/interruptible-universes</loc>
<changefreq>monthly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/explainers/value-engine</loc>
<changefreq>monthly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/explainers/gradia-guard</loc>
<changefreq>monthly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/what-a-benchmark-actually-is</loc>
<lastmod>2026-08-11</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/can-you-trust-an-ai-to-grade-an-ai</loc>
<lastmod>2026-08-11</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/the-benchmarks-are-broken</loc>
<lastmod>2026-08-12</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/evaluating-agents-is-a-different-discipline</loc>
<lastmod>2026-08-12</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/your-failures-are-your-training-data</loc>
<lastmod>2026-08-13</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/how-to-write-an-evaluation-that-holds-up</loc>
<lastmod>2026-08-13</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/the-statistics-of-evaluation</loc>
<lastmod>2026-08-14</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/red-team-your-rubric</loc>
<lastmod>2026-08-14</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/evaluation-doesnt-end-at-launch</loc>
<lastmod>2026-08-15</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/trust-that-doesnt-require-trusting-anyone</loc>
<lastmod>2026-08-15</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/six-months-in-is-it-working</loc>
<lastmod>2026-08-16</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/eight-workflows-eight-failure-maps</loc>
<lastmod>2026-08-16</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/the-evaluation-maturity-model</loc>
<lastmod>2026-08-17</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/how-to-run-a-model-bake-off</loc>
<lastmod>2026-08-17</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/what-to-ask-a-vendor-about-their-benchmark</loc>
<lastmod>2026-08-18</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/build-or-buy-your-evaluation</loc>
<lastmod>2026-08-18</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/rag-evaluation-is-two-problems</loc>
<lastmod>2026-08-19</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/evaluating-multi-agent-systems</loc>
<lastmod>2026-08-19</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/the-economics-of-evaluation</loc>
<lastmod>2026-08-20</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/human-annotation-that-doesnt-fall-apart</loc>
<lastmod>2026-08-20</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/synthetic-data-for-evaluation</loc>
<lastmod>2026-08-21</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/benchmarks-with-a-date-on-them</loc>
<lastmod>2026-08-21</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/evaluation-in-regulated-industries</loc>
<lastmod>2026-08-22</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/the-30-day-evaluation</loc>
<lastmod>2026-08-22</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/the-number-nobody-publishes</loc>
<lastmod>2026-08-23</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/an-empty-checklist-passes-every-audit</loc>
<lastmod>2026-08-23</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/test-the-grader-not-just-the-model</loc>
<lastmod>2026-08-24</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/your-traces-are-a-benchmark-you-havent-compiled</loc>
<lastmod>2026-08-24</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/when-the-simulated-user-leaks</loc>
<lastmod>2026-08-25</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/say-what-youre-predicting-before-you-know</loc>
<lastmod>2026-08-25</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/audit-the-passes-not-just-the-failures</loc>
<lastmod>2026-08-26</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/every-number-should-resolve-to-a-receipt</loc>
<lastmod>2026-08-26</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/seed-the-corruption</loc>
<lastmod>2026-08-27</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/the-grader-needs-a-grader</loc>
<lastmod>2026-08-27</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/an-unverified-oracle-is-not-an-answer-key</loc>
<lastmod>2026-08-28</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/silence-is-never-success</loc>
<lastmod>2026-08-28</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/a-final-state-benchmark-cannot-prove-the-world-changed</loc>
<lastmod>2026-08-29</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/a-benchmark-score-also-grades-the-harness</loc>
<lastmod>2026-08-29</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/the-scorekeeper-is-an-attack-surface</loc>
<lastmod>2026-08-30</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
<url>
<loc>https://www.gradiahq.com/blog/a-failure-map-is-not-a-training-plan</loc>
<lastmod>2026-08-30</lastmod>
<changefreq>yearly</changefreq>
<priority>0.6</priority>
</url>
</urlset>
