Codex power users say Astra burns $200 limits in days while Anthropic's cap is hard to hit
brandon_galang · x · 2026-09-27
Longtime Codex user Nicolas Zu aired complaints about OpenAI's $200 plan: Astra chews through limits in days even when he tries to optimize, the new 6 Sol model feels lazy and worse than Sol 5.6, and Sol 5.6 is a league below Opus 5.5 — while the $200 Anthropic plan is virtually impossible to exhaust. He still loves the Codex harness and app. Amplifier brandongalang adds he rarely hit limits on 5x in earlier generations but Astra drains even 20x, a contrast with OpenAI's efficiency messaging, though he remains optimistic about Dev Day.
More from coding & agent
- Coding agents show real agency by handing you a script when sandbox blocks them — jobergum · 2026-09-27
- Semantic emoji search with on-device bge-small beats LLM at $0 per search — maxleiter · 2026-09-27
- Why Is There No "Hardened Sandbox" Startup Yet for AI Agents? — MannyKayy · 2026-09-27
- Tell your coding agents to replace if-else with asserts: fewer bugs, less code — gabriberton · 2026-09-27
- Zuckerberg: most firms won't own frontier AI, but will build their own AI operating layer — rohanpaul_ai · 2026-09-27
- From 100M to 5B tokens a day: multi-agent swarms emerge as a new scaling axis — xeophon · 2026-09-27