Editorial Policy

1. Who Writes and Reviews

RankLLMs is founded and operated by Lucky Yaduvanshi, a software engineer and DevOps specialist, who is the lead researcher and final reviewer for all published material. Articles carry a named author, an original publication date, and a visible last-updated signal when the data inside has been refreshed.

2. Sourcing Rules

• Model specifications and pricing are refreshed through a scheduled sync from models.dev and OpenCode, with changes proposed for review. The sync preserves benchmark score fields and editorial text. Published benchmark results are not described as RankLLMs tests unless a page documents a first-party test.
• Every non-obvious factual claim in a data-heavy article is attributable: either linked inline or named in the text.
• Historical results are labeled with the exact model version and release date they belong to, so a 2024 score is never presented as a current one.
• Opinion and analysis are clearly framed as interpretation, and they are grounded in the data shown in the same article.

Fact Checking and Evidence Standards

Check factual claims against the linked source and its date, model version, benchmark definition, or offer terms. Published vendor results and third-party benchmark results must be attributed to their source; analysis of those results should not be described as a RankLLMs test. A first-party testing claim needs supporting methodology and results. If supporting evidence is unavailable, describe the finding as sourced analysis or state the uncertainty.

Use the correction process below to report a claim that conflicts with its source or lacks adequate evidence. Editorial independence does not mean that RankLLMs performed every cited benchmark.

3. Independence and Pay-to-Rank

No vendor can buy a position, score, or review outcome on RankLLMs, and providers do not preview rankings before publication. The Site currently carries Google AdSense advertising. Affiliate or referral relationships are disclosed clearly and near the related link or recommendation; neither advertising nor referral revenue determines a score or editorial conclusion. See the disclaimer for details.

4. Open Source and Auditability

The scoring project is published under the MIT license in the rankllms-engine repository. Readers can inspect the implementation; this does not mean RankLLMs independently reproduces every upstream benchmark.

5. AI Assistance Disclosure

AI tools may assist with research organization, drafting, code, or data formatting. A human editor is responsible for published conclusions and checks material claims against their sources. AI-generated output is not treated as benchmark evidence or published as an unreviewed verdict. Where substantial AI assistance shapes an article's analysis, that assistance is disclosed in the article.

6. Corrections and Updates

• Factual errors are corrected as soon as verified, and the page notes that it was corrected.
• Data refreshes (new benchmark data, pricing changes) update the page and bump the visible last-verified date; the article's original publication date stays fixed.
• Material changes of conclusion - for example, a revised verdict after new results - are described in the page itself, not silently edited away.
• Report errors through the contact page. Verified material corrections are dated on the affected page.

7. Sponsored and Deal Content

Offer and review pages may link to a provider or referral program. If RankLLMs receives compensation or another material benefit, the disclosure appears clearly near the relevant recommendation or link so readers can evaluate it in context. A disclosure on a general policy page alone is not a substitute for a page-level disclosure. Sponsored cards carry a visible "Sponsored" tag, deal pages show an affiliate disclosure box, and individual articles claim commissions only when their frontmatter sets `affiliate: true`. Offers can expire or change; confirm their current terms with the provider.