Methodology hub
Research
Source-aware research notes and infrastructure explainers built on the verified WebmasterID Models entity graph. Each page is internally linked to the underlying model, provider, pricing, and source records so every assertion can be checked against the data.
Research guides
7 guides
Research guides
Choosing an AI model: a verified-data approach
A practical, source-aware framework for selecting an AI model — covering provider, pricing, context window, output limits, modality, lifecycle, and reliability signals.
Last updated: 2026-05-21
Read guide →
Research guides
API pricing methodology: what the rows on /pricing mean
How WebmasterID Models tracks AI API pricing — input tokens, output tokens, cache write windows, cache reads, per-hour cache storage, batch tiers — and why provider pricing cannot always be normalised into a single number.
Last updated: 2026-05-21
Read guide →
Research guides
Context windows in practice
What a context window actually means for production workloads, why a million-token window on one provider is not the same as a million-token window on another, and how WebmasterID Models verifies the number.
Last updated: 2026-05-21
Read guide →
Research guides
Output token limits and what they constrain
Why max output tokens is a separate dimension from the context window, how it shapes long-form generation, agentic workflows, and structured output, and which models publish what.
Last updated: 2026-05-21
Read guide →
Research guides
AI provider status monitoring: vendor signals vs independent probes
How WebmasterID Models separates vendor-reported status, independent HTTP probes, and computed uptime windows — and why no uptime percentage is published without durable observations.
Last updated: 2026-05-21
Read guide →
Research guides
Why benchmark scores need source discipline
Benchmark scores are useful only when their provenance, dataset, and prompt protocol are documented. This page explains why WebmasterID Models does not republish unsourced provider-reported scores.
Last updated: 2026-05-21
Read guide →
Research guides
AI inference infrastructure — fields, gaps, and roadmap
Regions, cloud availability, status feeds, batching, caching, rate limits, throughput — the infrastructure dimensions a verified catalogue cares about, and which fields remain unverified today.
Last updated: 2026-05-21
Read guide →
Verification policy