Fable 5 vs. 5.1: Build Efficiency and Cost Comparison
Nate Herk | AI Automationgo watch the original →
the gist
A side-by-side comparison shows Fable 5 is more cost-effective and faster for building a visual automation app, despite Fable 5.1 scoring slightly higher in subjective UI quality.
The Breakthrough
Fable 5 outperformed Fable 5.1 in build efficiency, completing the same visual automation application in 12 hours for $500, compared to 36 hours and $1,200 for the 5.1 version.
What Actually Worked
- The author utilized a delegation-heavy prompt strategy, forcing the agents to act as orchestrators rather than individual engineers. The prompt explicitly mandated: "You own strategy, planning, delegation, sequencing, quality standards, and final acceptance. Do not personally perform engineering research, debugging, testing, or visual design production."
- The prompt enforced a strict worker hierarchy, requiring the use of Claude 3.5 Sonnet for implementation, research, and testing, while reserving Claude 3.5 Opus for architecture, design direction, and difficult problem-solving.
- Fable 5 achieved higher efficiency by relying on Sonnet for 80% of its task execution, whereas Fable 5.1 utilized Opus for 57% of its workload, significantly increasing the total cost.
- The author implemented a blind testing protocol using an external reviewer (Codeex) to evaluate the resulting applications on UI hierarchy, ease of use, and workflow authoring without knowing which agent built which version.
Before / After
- Fable 5: 12 hours build time, $500 cost, 260,000 tokens context usage.
- Fable 5.1: 36 hours build time, $1,200 cost, 404,000 tokens context usage.
- Blind Test Score: Fable 5.1 scored 9.1/10, while Fable 5 scored 8.4/10.
Context
The author tested two versions of the Fable AI agent to build "Ops Flow," a local-first visual automation studio for incident response workflows. The experiment aimed to determine if the newer 5.1 model provided enough value to justify the significant increase in cost and time compared to the previous version. While 5.1 produced a more polished UI, the author concluded that the marginal improvements did not justify the $700 price difference, declaring Fable 5 the winner for production efficiency.
Content References
{"type": "tool", "title": "Fable", "context": "reviewed"}, {"type": "tool", "title": "Claude 3.5 Sonnet", "context": "mentioned"}, {"type": "tool", "title": "Claude 3.5 Opus", "context": "mentioned"}