AI plan comparisons
The headline price of an AI plan is the easy part. The hard part — the part that decides whether a plan is actually worth it — is the limit: how much you can do before it meters out, and how honestly the vendor tells you. Most explainers go stale within a month of publishing. These don't: each one quotes the vendor's own wording, prices everything in canonical USD, and is re-checked against the live ledger every day. Pick the comparison you need.
The comparisons
- Claude Pro vs Max Exact prices, what each plan includes, and what Anthropic's 5x / 20x usage tiers actually mean — quoted verbatim. Finding: Anthropic never publishes the base number, so "20x" is twenty times an unknown. Pro ($17 annual / $20 monthly) includes Claude Code.
- ChatGPT Plus vs Pro Plus ($20) vs Pro (from $100/mo): what the jump buys, and what OpenAI's 5x / 20x tiers really mean. Finding: A $20 → $100 cliff with no tier in between, and ads now reach the Free and Go tiers.
- Cursor vs GitHub Copilot Plan-by-plan prices and limits for the two most popular AI coding subscriptions — and the two things buyers miss. Finding: Copilot is cheaper and publishes real credit numbers, but its paid sign-ups are paused; Cursor is buyable but prints no numeric quota.
- Google Gemini plan limits explained Free, AI Plus ($7.99), AI Pro ($19.99), and AI Ultra ($99.99 / $199.99) — the one major that actually publishes its limit mechanics. Finding: Gemini publishes the machinery — a 5-hour refresh, a weekly cap, and Flow credits as a hard number — where rivals give only multipliers.
- Which AI coding plan has the best limits? Claude Code, Cursor, Copilot, Codex, Gemini, Devin and Mistral coding plans ranked by how their usage limits actually work. Finding: Almost nobody publishes a comparable coding number. Copilot has hard figures but paused sign-ups; Cursor is the most honest buyable plan.
Why these stay accurate
AI plan limits and prices are among the most-adjusted figures in software, and they usually change quietly — a blog refresh here, a tier rename there. QuotaLedger snapshots every tracked vendor's pricing page daily, preserves the wording verbatim, and logs any change with a timestamp. So when one of these comparisons says a plan costs $X or includes "5x usage," that reflects what the vendor said on the date stamped at the top — not whatever was true when the page was first written.