Fable 5.1 review: Burns through 20x usage in 52 minutes on single prompt

tetsuoai · x · 2026-09-02

The user reported testing Fable 5.1, noting that it consumed 20x usage credits in just 52 minutes from a single prompt without returning code. They warned that the model spawns many agents and is expensive to run, suggesting gating it to a single agent. In a follow-up retry with limited agents and restored quota, the author praised the model as extremely good.

Related event: Fable 5.1 tops benchmarks but running costs surge, sparking smarter-or-pricier debate(23 posts)→

Original post →

More from Models

Models channel →