Opus may cost more than Fable on hard debugging when retries pile up
doodlestein · x · 2026-07-27
The poster argues that Opus is not necessarily cheaper for hard debugging or complex implementation work.
If a model has to retry the same task multiple times because it keeps making mistakes, the total cost and time can end up higher than using a model like Fable that gets it right on the first pass. A reply adds that Opus 5 often surfaces a long list of mistakes after the session, while Fable tends to think more before acting and avoids those errors up front.
More from coding & agent
- He benchmarks new models by having them build their own copilot — $28 of compute — bucolucas · 2026-10-27
- Should agents be allowed to change plans mid-task? Reddit debates the drift boundary — wistfulcracker84 · 2026-10-27
- HeyGen + Canva MCP turns slide speaker notes into an avatar presentation video — HeyGen · 2026-10-27
- Local beats cloud: Qwen 3.8 FN handles private student data as well as frontier models — dr_lm · 2026-10-27
- Cloudflare Computer adds agent interruption recovery and approval pauses in v0.7.0 — irvinebroque · 2026-10-27
- LLM Tails cuts agent input tokens 80% by making agents read before they write — savadks1818 · 2026-10-27