ChatGPT vs Claude vs Gemini vs Grok — the one-page card that picks your $20 AI plan by receipts, not vibes
The companion to the video. On August 13, 2026 I ran the same 5 real jobs through the exact frontier model each consumer plan unlocks — GPT-5.6 Sol, Claude Fable 5, Gemini 3.1 Pro, and day-one Grok 4.6 — with machine grading wherever possible: a bug-fix fuzzed against 500 randomized cases, 12 messy invoice emails checked field-by-field against planted ground truth (60 fields), 5 memos buried in a novel-length document, 10 exact-answer math problems, and one awkward landlord email judged blind by the other vendors' models with self-votes deleted. Result: 16 perfect scores and a photo-finish on the essay — the intelligence race is a tie, so your $20 buys limits, features, and the free tier underneath. INSIDE: (1) FIND YOUR ROW — five personas, five verdicts (talkers pay nobody: free ChatGPT text went unlimited Aug 6; coders take Claude Pro — the only $20 plan with an agent in your own terminal; makers and Gmail/Drive households take Google AI Pro's bundle with 1,000 Flow video credits, Jules, 5TB storage and YouTube Premium Lite; X-natives take SuperGrok at its REAL price). (2) THE REAL-LIMITS SHEET — all four plans' top model, app context, standout, and catch on one page, checked against the four official pricing pages the same day, including the headline catch: there is no $20 Grok — SuperGrok is $30, a 50% premium. (3) THE 5-MINUTE SELF-TEST — three copy-paste prompts (real-task probe, judgment probe, limit probe) you run in the FREE tiers with your own work before paying anyone. (4) THE FULL SCOREBOARD with method notes and the latency receipt (ChatGPT ~28s vs Gemini ~69s on identical math — same answers, 2.4x the wait). (5) SIX MONEY RULES, including the quarterly re-decide play that keeps your $240/yr working for you. Independent kit from Hyperautomation Labs — not affiliated with OpenAI, Anthropic, Google, or xAI.
Free. No spam. Unsubscribe anytime.