Block Uses Fable to Orchestrate Migrations
Block reports using Claude Fable 5 to coordinate multi-model workflows for large-scale code migrations, reducing reliance on single frontier models.
ImportanceLocalEvidenceE2 unreplicatedWrite-upQuick
Bradley Axen, Head of AI Capabilities at Block, stated that the company is using Claude Fable 5 as an "orchestrator" for large-scale code migrations involving thousands of pull requests.
In this workflow, Fable handles high-level design tasks such as data models and API specifications, while directing smaller, more cost-effective models like Opus or Sonnet to perform individual file edits and run tests. This approach aims to balance the planning capabilities of frontier intelligence with the execution efficiency of smaller models.
The case illustrates a shift in enterprise AI adoption from "always use the strongest model" to "hierarchical model orchestration." Axen noted that this is the first time they found that not every task requires the highest-tier frontier model. Specific cost savings percentages and quality impacts were not independently verified.