One month with ChatGPT's $500 Pro plan: slow models, tiny limits, stick to local AI
croakyossum7 · reddit · 2026-10-27
A college developer's detailed one-month review of ChatGPT Pro ($500/mo): not worth it — invest in local hardware instead.
- GPT-6.1 Sol: reasonably capable but painfully slow (30 tok/s) with poor instruction-following — he had to ask six times to remove a rendering engine, and forked code never actually landed in his repo; Codex Cloud made little real progress in 3 weeks.
- Astra: much stronger but limits are appalling — 7 hours of continuous use vs Sol's 2-4 days; UltraFast drains to 0% in about an hour. At twice Claude's top plan price, he must choose between a weak midrange model that lasts the week and a frontier model that dies in a day, while Claude's best model IS its midrange.
- Reliability: for six days half of responses failed with "Response could not load"; support's fix was to use the slower GPT-6 Pro; bugs reported months ago on Plus remain unfixed.
- Local alternative: Qwen 3.8 Flash Next on an RTX 5080 hits 40 tok/s (Q3), faster than Sol, free and unlimited; Qwen 4 coming.
Verdict: heavy users should pick a cheaper subscription and buy hardware.
More from Infra
- Critical pre-boot RCE in EDK II PXE driver: one crafted packet hijacks servers before Secure Boot — cyb3rops · 2026-10-28
- Screenless M1 Airs at $30 per 1,000 multicore points spark cheap Mac cluster idea — vincent_koc · 2026-10-28
- RTX 3080 Tensor Core unlock claims 80-87% TFLOPS boost, RTX 50 series blocked — Revolutionalredstone · 2026-10-28
- CXMT to start small-scale HBM3E production within two months, yield at 20% — zephyr_z9 · 2026-10-28
- Video Transcription Costs Under 1 Yuan per Hour, CPU Free but Error-Prone — huangyun_122 · 2026-10-28
- A forgotten supercomputer design: bare dies, gold post interconnects, 3D stacking, immersion cooling — jwt0625 · 2026-10-28