Founder & Lead AI Benchmark Researcher
Lucky Yaduvanshi
Software Engineer, DevOps Specialist, and AI Researcher. I built and run RankLLMs - the independent AI model leaderboard - and personally verify every score, price, and verdict published here.
What I do at RankLLMs
- 01
Own the dataset
Every one of the 85 model scorecards - benchmark scores, context windows, inference speeds, and blended API pricing - is sourced from official model cards and vendor pricing pages, then checked and dated before it goes live.
- 02
Field-test CLI coding agents
Hands-on terminal work with tools like Claude Code, Gemini CLI, and Copilot CLI: real repos, real issues, real token bills. The agent reviews on this site come from those runs.
- 03
Track the market weekly
The top 50 models are re-reviewed at least weekly. When a provider ships a new checkpoint or changes pricing, the leaderboard reflects it - and the change is dated.
- 04
Write the verdicts
The comparison guides and cost-per-solved-task math on this site are written from the same verified dataset the leaderboard uses, so a verdict and a ranking never disagree.
How my work is verified
Every scorecard shows when its data was last verified, and historical results are labeled with the exact model version they belong to. The full sourcing and scoring formula is public on the methodology page, and errors are corrected in public - editorial policy explains how.
Independence
RankLLMs is an independent publication, not owned by or affiliated with any AI provider. Rankings are never sold: labs cannot preview, influence, or purchase placement on the leaderboard. The site is funded by me, and any future advertising or affiliate relationship will be disclosed where it appears.
Latest articles
View all (38) →
Publishing on RankLLMs since August 2026 • Questions or corrections? Get in touch.