Users criticized Fable AI's double model fallback to Haiku after Opus safety filters, complaining they pay premium prices for inferior unreliable outputs from flawed classifiers.
Based on 6 visible X reactions from 39 accounts; directional sample.
Ask a question below.
Published answers will appear here.
@willccbb Wait for Opus 5. This is gonna get significantly worse. Anthropic has leaned very hard into safety, but their classifiers are not good enough to correctly “classify”.
2:23 AM · Jul 15, 2026@willccbb and you'll be charged fable pricing and opus pricing for something that sonnet will do
2:45 AM · Jul 15, 2026@willccbb haiku 3 for slides is rough. paying opus prices to get served a classifier's leftovers
2:03 AM · Jul 15, 2026@willccbb Drop @AnthropicAI now and come be happy on Codex 😎
4:10 AM · Jul 15, 2026wow double fallback. asked fable to help with some slides. they’re gonna route me to haiku 3 next
1:53 AM · Jul 15, 2026@willccbb You are attempting to talk about frontier research 🔴
2:04 AM · Jul 15, 2026wow double fallback. asked fable to help with some slides. they’re gonna route me to haiku 3 next
1:53 AM · Jul 15, 2026@willccbb You are attempting to talk about frontier research 🔴
2:04 AM · Jul 15, 2026Users criticized Fable AI's double model fallback to Haiku after Opus safety filters, complaining they pay premium prices for inferior unreliable outputs from flawed classifiers.
Based on 6 visible X reactions from 39 accounts; directional sample.
Ask a question below.
Published answers will appear here.