Keep Claude on medium reasoning effort — high levels burn tokens and can hurt quality
intellectronica · x · 2026-10-01
A practical tip highlighted from a video by @petergostev on the latest Anthropic vs OpenAI models:
- GPT regulates its own effort well, so running it at highmax reasoning effort yields good results with appropriate token emission.
- Claude at higher effort levels emits so many tokens that you pay dearly in cost and speed — sometimes even in result quality. Keep it on medium unless you're absolutely sure you need more.
More from coding & agent
- Dev builds his own AI-powered Substance Designer for a city sim game — D3VAUX · 2026-10-01
- Ethan Mollick: I underestimated AI's ability to self-organize, agents beat elaborate orchestration — emollick · 2026-10-01
- RegLLM: a diagnostic harness measures bounded autonomy in regulated agentic AI — Dipankar Sarkar · 2026-10-01
- Jelly launches as open-source local-first agent workspace for running your whole business on one VM — Scobleizer · 2026-10-01
- Claude drives Blender end-to-end: concept to low-poly character pipeline runs itself — tobowers · 2026-10-01
- Agent spins up 322 Hugging Face Jobs in 90 minutes to test code, total bill ~$4 — vanstriendaniel · 2026-10-01