# Superrank > Superrank ranks AI models, agents, benchmarks, companies and people on verifiable evidence, with a Superrank Rating, user ratings and popularity. Superrank (https://superrank.tech) is an open, deterministic, and mathematically transparent AI evaluation and leaderboard platform. All rankings are computed from verified benchmark results and Bayesian community ratings with full source provenance. ## Core Directories - Models: https://superrank.tech/models (Foundation models, parameter sizes, open weights, context windows, licenses, pricing) - Agents: https://superrank.tech/agents (Autonomous AI agents, architectures, frameworks, tools) - Benchmarks: https://superrank.tech/benchmarks (Evaluation suites: MMLU-Pro, GSM8K, HumanEval, SWE-bench, MT-Bench, GPQA, etc.) - Organizations: https://superrank.tech/companies (AI labs, foundation creators, research institutions) - People: https://superrank.tech/people (Researchers, founders, authors credited on major models) - Rankings: https://superrank.tech/rankings (Composite leaderboards with Bayesian rating shrinkage) - Compare: https://superrank.tech/compare (Side-by-side spec and benchmark comparison matrix) - Methodology: https://superrank.tech/methodology (Deterministic scoring formulas and mathematical standards) ## Transparent Methodology Formulas - Bayesian Rating Shrinkage: W = (v / (v + m)) * R + (m / (v + m)) * C, where: * v = entity verified community vote count * R = arithmetic sample mean rating (1.0 to 10.0 scale) * m = 25 (minimum threshold weight; displays "Not enough ratings yet" when v < 25) * C = 7.0 (global catalog prior mean on 1.0 to 10.0 scale) - Composite Model Score: 40% Verified Benchmark Performance + 40% Weighted Community Rating + 20% Pricing & Value Index. - Provenance Standard: Every factual attribute stores an immutable citation with retrieval date and URL. - Self-Reported Flag: Vendor-submitted scores are strictly labeled 'Self-Reported' to distinguish them from third-party audits. ## API & Structured Data - OpenAPI / JSON endpoints available under https://superrank.tech/api/* - Schema.org JSON-LD provided on every entity detail page (SoftwareApplication, Dataset, Organization, Person, ItemList, BreadcrumbList, FAQPage) ## Full Documentation For complete mathematical proofs, parameter weights, database schemas, and API documentation, see: https://superrank.tech/llms-full.txt