AI coding tiers become governance points

OpenAI has inserted a new $100-a-month ChatGPT tier into the middle of its coding lineup, turning price plans into a way to sell control over usage. (techcrunch.com) OpenAI announced the plan on April 9, 2026, after months with a gap between $20-a-month Plus and $200-a-month Pro. The new tier gives five times more Codex usage than Plus, and OpenAI offered a temporary 10 times launch promotion through May 31, according to coverage citing the company. (techcrunch.com) (thenextweb.com) Codex is OpenAI’s coding tool inside ChatGPT, and OpenAI told TechCrunch that Plus and the new $100 Pro tier are both aimed at daily coding use. TechCrunch also reported that OpenAI said more than 3 million people use Codex each week, up fivefold in three months. (techcrunch.com) The competition is no longer just about which model writes better code. It is also about how much coding work a provider allows before rate limits hit, how many parallel projects a plan can support, and what a company promises about where requests are handled. (techcrunch.com) (opencode.ai) OpenCode’s Go plan shows how those promises are moving into product design. Its documentation says the beta service costs $5 for the first month and $10 a month after that, is built for international users, and routes models through the United States, European Union, and Singapore for “stable global access.” (opencode.ai) OpenCode also publishes hard usage ceilings in dollar terms instead of vague “fair use” language. Its Go docs list a 5-hour limit of $12 in usage, a weekly limit of $30, and a monthly limit of $60, with request counts varying by model cost. (opencode.ai) Mistral is making a related argument from the European side of the market. In a playbook published April 7, 2026, the French company said Europe should build “homegrown AI infrastructure,” grow local talent, and reduce reliance on foreign providers. (mistral.ai) That pitch lines up with a buyer concern that has moved from procurement paperwork into the product itself: where the model runs, which laws govern the data, and whether a team can keep work inside a region. OpenCode’s regional hosting menu and Mistral’s sovereignty push both package geography as a feature, not a footnote. (opencode.ai) (mistral.ai) OpenAI’s new tier also lands in a direct pricing fight with Anthropic. TechCrunch reported that OpenAI described the $100 plan as a challenge to Anthropic’s long-standing $100 Claude option, while The Next Web said the move also raises pressure on Google in coding subscriptions. (techcrunch.com) (thenextweb.com) The result is a market where the monthly bill now signals more than model access. It also signals how much autonomy, capacity, and jurisdiction a coding team can buy. (techcrunch.com) (opencode.ai) (mistral.ai)