Glossary · Term

Fable 5

← all terms

Definition

Plain language

One of the AI models used in these experiments.

As stated in the literature

A Claude model variant referred to in the scripts as claude-fable-5; the top-scoring agent on FM-Bench's twenty-year solo track despite being among the cheapest in token spend.

Also called: claude-fable-5

Why it matters: It shows that a model can be deliberately tuned toward unusual, varied answers, which matters when you want creativity or breadth rather than the safest common response.

For example, asked to name a fruit, Fable 5 might answer "quince" while most other models say "apple."

Heard on the show

“The best of them, claude-fable-5, scores ninety point nine four.”
Episode 245 — Fifteen Models Ran Football Clubs for Twenty Years, and Size Didn't Decide It

Mentioned in 23 episodes

  1. 245
    Fifteen Models Ran Football Clubs for Twenty Years, and Size Didn't Decide It
  2. 240
    Frontier Models Designed Follow-Ups To Fraudulent Papers 93% Of The Time
  3. 229
    One Word Flips a Chatbot From Backbone to Yes-Man
  4. 219
    Forty-Four AI Models, One Word, And The Newest Ones Conform Most
  5. 206
    How Four-Second Clips Become Hours of Playable AI Soccer
  6. 205
    The Same AI, Two Labels: How the Pitch Beat the Product in 162 Sessions
  7. 204
    The Length Estimate Hiding Inside a Word-by-Word Model
  8. 203
    The Thought a Model Doesn't Say — and the Lens That Reads It
  9. 201
    One in Four NeurIPS Papers Cites a Reference That Doesn't Exist
  10. 200
    The One Mechanism That Turns Twenty AI Clones Into an Actual Team
  11. 199
    Finding a Model's Hidden Behaviors Without Knowing What You're Looking For
  12. 198
    The Model That Knows the Answer and Can't Say It
  13. 197
    Twin Problems Suggest AI Reasoning Gains Are Mostly Better Fact Recall
  14. 196
    AI Agents Reached Opposite Conclusions From the Same Data — and Passed Review
  15. 195
    Why 'Be Careful' Does Nothing for AI Coding Agents, and What Does
  16. 190
    The Skill Every AI Manager Is Missing: Handing Out Exactly the Right Keys
  17. 133
    How MiniMax Turned a Reward-Hacking Disaster Into Olympiad Gold
  18. 132
    The Agent Failed — But Did the Instructions Deserve to Be Followed?
  19. 131
    Why Autonomous Research Agents Forget Their Own Lessons, and Arbor's Fix
  20. 130
    Why AI Agents Coordinate Better Through a Shared Board Than a Boss
  21. 129
    How a Crowd of Anonymous AI Agents Broke a 40-Year Math Record
  22. 128
    How a Model Can Earn Full Reward and Still Resist Training
  23. 127
    What Diffusion Language Models Were Missing: A Map, Not an Algorithm

Related terms

agent Claude FM-Bench token