User finds Claude Haiku keeps botching simple math question drafting, falls back to Opus
joshh0444 · reddit · 2026-09-11
A Reddit user tested the common advice of delegating grunt work to Claude Haiku and hit repeated failures: while drafting algebra multiple-choice questions via Claude Code, Haiku solved its own problems incorrectly three times in a row, each time catching the mismatch and rewriting the question instead of delivering. The user ended up regenerating everything with Opus — which still needed refinements — and questions whether small models are really suitable even for simple task drafting.
More from Models
- Dev Claims DeepSeek v4.1 Flash "Way Better" Than Gemini Flash Models — gaganghotra_ · 2026-09-12
- Altman Talks GPT7 in New Video Clip — Popular_Slip_5311 · 2026-09-12
- LinkedIn user claims GPT-6 built a pixel-perfect Figma design system in 3 hours — AIandDesign · 2026-09-12
- Report: Google Achieved Recursive Self-Improvement in Next-Gen AI Model, Launching Oct 5 — Dr_Singularity · 2026-09-12
- Leaked Codenames Suggest OpenAI Model Tiers Map to Claude's Opus/Sonnet/Haiku — haider1 · 2026-09-12
- Eric Xing on "open source": open weights is a house with no blueprints — YiMaTweets · 2026-09-12