Heidelberg AI News

Video · David Ondrej ·

My $1,000 Fable 5.2 Workflow (just copy me)

In short

The video explains a coding workflow that combines Fable 5.2 for difficult planning and reasoning, Opus 5.5 for faster execution, and Gemini 4 Carbon for independent reviews. It matters because Fable is highly capable but consumes limits quickly, so model selection and task-based routing reduce cost and latency.

Key points

  • The creator sent over 1.6K prompts across 20+ hours and hit usage limits 53 times in 3 days.
  • Fable 5.2 is best for ambiguous, deeply technical, and high-value tasks, while Opus 5.5 handles clear execution faster.
  • A practical workflow is to start a thread with Fable, then switch to Opus once the plan is clear.
  • Use Raycast snippets and reusable skills; 34% of the creator's prompts came from 31 snippets.
  • Use Fable at high reasoning effort by default, reserving max for especially valuable investigations.
  • Gemini 4 Carbon is used as an independent review agent because its architecture and training differ from the Claude-family models.
  • The creator runs 35 agent loops, with 72% using Opus 5.5 and 3% using Fable 5.2.

Written by an AI model from the video’s captions and description.

In this video David Ondrej shares what he learned from spending over $1,000 on Fable 5.2 via Claude Code subscriptions, and how he uses it with the new Gemini 4 Carbon.

Watch on YouTube The opening of the article at youtube.com.

In this story

A tool or model opens in the Heidelberg AI Curriculum: what it does, what it costs, how to start. A repo opens on GitHub.

Related stories

All stories

VideoWes Roth

Claude JUST got MYTHOS...

The video argues that Claude is evolving into an AI operating-system layer that coordinates office tools, meetings, data, software, and third-party apps rather than replacing each app directly. It also explains how eligible independent researchers, open-source developers, and vulnerability hunters may apply for Mythos 5.2 through Anthropic’s Cyber Verification Program, while challenging the interpretation of the MIT study claiming 95% of enterprise AI projects produce no ROI.