I Rebuilt All 7 Claude Fable 5 Demos Anthropic Showed 4.2M People — Here's the Scoreboard
The official Claude account posted seven Claude Fable 5 demos to 4.2 million people — a goldfish pond, a Boeing 747, an infinite gothic city drowning in the ocean. Most people scrolled. I took all seven prompts, opened a blank editor, and rebuilt every single one myself, blind — then graded each one: ONE-SHOT, FIXED, or PARTIAL. Three ran clean on the first try. Two needed a steadier hand. One was honest about its broken physics. And one was never a one-shot at all. Then, while I was testing, the U.S. government stepped in and suspended the model. Here's the full scoreboard. Comment SEVEN and I'll send you the free 7-Demo Scorecard — all seven prompts word-for-word so you can run the exact same test, my grade on each, what broke, every source linked, and the one trick (the self-checking loop from the 747) that mattered most. ⏱ TIMESTAMPS 0:00 7 demos, 4.2M views — I rebuilt them all 1:14 The rules of the test 2:22 Demo 1 — Goldfish pond (ONE-SHOT) 3:34 Demo 2 — Suminagashi ink fluid (FIXED) 5:06 Demo 3 — Boeing 747, the self-verifying loop (EARNED) 6:59 Demo 4 — Drowned gothic shader (ONE-SHOT) 8:14 Demo 5 — The wooden house benchmark (FIXED) 9:37 Demo 6 — The poseable ragdoll (REAL · paper legs) 11:02 Demo 7 — The soccer trail (PARTIAL) 12:33 The scoreboard — and what every demo has in common 14:06 The twist — the government pulled the model 15:24 The free guide CREDITS — these builds belong to their builders; the reproductions are my own test: • @midori_tatsuta — the goldfish app • @hayashimon1 — the suminagashi ink fluid • @victormustar (Hugging Face) — the Boeing 747 benchmark • Ethan Mollick (@emollick) — the twigl gothic-city shader • @Angaisb_ — the functional wooden-house benchmark • @jasonkneen — the poseable soft-body character • @measure_plan — the soccer computer-vision project • Original roundup: the @claudeai account · Anthropic: anthropic.com/news/claude-fable-5-mythos-5 Facts as of June 15, 2026: Claude Fable 5 is the public version of Anthropic's Mythos-class model, included free on Pro, Max, Team and seat-based Enterprise plans through June 22, then metered on usage credits from June 23 (API list price $10 input / $50 output per million tokens — about 2x Opus 4.8). On June 13, 2026 the U.S. government issued an export-control directive suspending Fable 5 and Mythos 5 for foreign nationals (anthropic.com/news/fable-mythos-access). 📸 Bite-size AI drops every day on Instagram: instagram.com/hyperautomationlabs ▶ Full deep dives here on YouTube — subscribe: / @hyperautomationlabs1045 #ClaudeFable5 #Anthropic #AI #ClaudeCode #AIcoding

Stop Prompting Claude. Use Karpathy's Method Instead.

We let AI buy a robot and a car, it does exactly what experts warned.

Why AI Can't Count the R's in "Strawberry" — 7 Mental Models (Karpathy's Masterclass, Simplified)

What do AI agents do when humans aren’t watching? - BBC World Service

Hands on: Fable 5 makes GPT 5.5 feel like a "toy"

The Head of Claude Code Just Revealed His Exact Workflow (Steal It)

I Made Opus 4.8 and Fable 5 Build the Same App (RAW RESULTS)

Finally. Agent Loops Clearly Explained.

The $15,000 AI Bill. Your $20 Subscription is a DELUSION

Claude Design Just Killed the Design-to-Code Handoff

Learn 97% of Claude in Under 16 Minutes

Backend web development - a complete overview

GLM 5.2 vs Claude was INSANE.

Can an open-source AI really compete with Claude 3.5 Sonnet?

Claude Fable 5 Made This Entire Video By Itself.

Claude Fable 5 is INSANE.

How Much Money my Indie Game Made (After 1 month)

I Think They Are Lying To You

The New AI prompting Method Everyone Is Talking About: Loops (And How To Use Them)

