aiminute. ← All AI news
Industry AI Minute Newsroom 2026-08-11

OpenAI's business plan kept hitting a five-hour wall, so it built a $125 seat to go over it

OpenAI's business plan kept hitting a five-hour wall, so it built a $125 seat to go over it

OpenAI announced Premium seats for ChatGPT Business on 10 August: $125 per user per month, or $100 with annual billing, against $25 for a standard seat ($20 annually). The money buys roughly five times the usage of a standard seat and, more to the point, removes the five-hour usage limit that had been cutting off Codex and Workspace Agent runs part-way through a job. Workspaces can mix the two seat types, so a team can buy the expensive one only for the people who keep hitting the ceiling. Early sign-ups get $100 in workspace credits per Premium seat, up to $500 for five seats, and there is a waitlist open until 20 August.

Why it mattersA fivefold price step aimed at the people who use a product most is what a company does when its heaviest users cost more than they pay. Agents are the reason. A chat session ends when you stop typing; an agent left to work through a task keeps consuming tokens for hours, whether anyone is watching or not. This is seat-based software pricing meeting a cost that scales with machine time rather than headcount — and OpenAI's answer, announced while it courts investors ahead of a possible listing, is to charge for the hours instead of pretending they are free. Anyone budgeting for AI tools should read it as a signal that flat per-user pricing will not survive contact with agents.
#AI Agents

✓ Verified · 3 sources

WhatsApp X Telegram
Read in the app — free, in 9 languages

Related stories

AI server makers are raising prices on deals already signed, FT says.
2026-11-22
OpenRouter dropped three AI providers after a bait-prompt leak test.
2026-11-22
Researchers say they broke out of Meta's Muse agent and leaked chats.
2026-11-22
Amazon is close to buying AI startup Decart for about $7.5 billion.
2026-11-22
Alibaba says it is training its next Qwen model on its own chips.
2026-11-21