I had Fable 5 for exactly 24 hours. In that time it hit 6.4% of my total model usage – across all models, over months.<\/p>\n
That’s not a complaint. That’s context.<\/p>\n
I didn’t run it for routine tasks. I pulled up projects that had been stuck – things where Sonnet did the work but something wasn’t adding up. Wrong assumptions, bad architecture decisions, edge cases nobody caught. Fable’s job was to find the problem and fix it.<\/p>\n
That kind of session burns tokens. You’re not generating content, you’ really thinking out loud with a model that actually keeps up. Long context, multiple files open, back and forth until it clicks. Of course the quota went fast.<\/p>\n
Fable held more in context without losing the thread. On complex debugging sessions – the kind where you have to trace a problem across several layers – it stayed accurate longer and needed less hand-holding.<\/p>\n
The 3.2M output tokens in the chart tell the story. I was in long sessions, not quick queries.<\/p>\n
I would have liked a week. Specifically to test the mixed setup – Fable for planning and root cause analysis, Sonnet or Haiku for execution. Keep the expensive model on the hard parts, offload the rest.<\/p>\n
That experiment didn’t happen. Fable was gone before I got there.<\/p>","protected":false},"excerpt":{"rendered":"