OpenAI just plugged the $180 gap in its pricing ladder, and the move tells you exactly who's winning the coding agent war.
The Summary
- OpenAI launched a $100/month ChatGPT Pro tier, sitting between the $20 Plus plan and the $200 Pro tier that already existed.
- The new plan offers 5x more Codex usage than Plus, directly targeting developers and "vibe coders" who build with AI.
- This is a direct shot at Anthropic's Claude Code Max tier, which costs the same $100/month.
- OpenAI is simultaneously rebalancing Plus tier limits to favor "more sessions throughout the week, rather than longer sessions in a single day," per VentureBeat.
The Signal
OpenAI now runs six pricing tiers for individuals alone: free, $8 Go, $20 Plus, $100 Pro, and $200 Pro, plus separate Edu, Business, and Enterprise tracks. That's not product strategy. That's price discovery in real time. The company is learning what developers will actually pay for agent-driven coding tools, and the answer came from users themselves. Sam Altman said the $100 tier launched "by very popular demand" after seeing how much love Codex was getting.
The math is clean: $20 times five equals $100, and you get five times the Codex usage limits. But the real story is competitive positioning. Anthropic's Claude Code tool has been pulling developers, and OpenAI needed a price-matched counterpunch. The $100 tier isn't just about revenue. It's about keeping the developer base inside the OpenAI ecosystem before they get comfortable elsewhere.
The rebalancing of the Plus tier is equally telling. OpenAI is pushing Plus users toward shorter, more frequent sessions instead of marathon coding blocks. That's capacity management dressed up as product tuning. Translation: demand for coding agents is outstripping what the current infrastructure can serve at $20/month, so they're rationing the Plus tier while upselling power users to $100.
What matters here is the category definition happening in public. "Vibe coder" is now a real user segment with real pricing tiers. These aren't hobbyists. They're people building production software by talking to models, and they'll pay $1,200 a year for the privilege. The agent economy isn't theoretical anymore. It has price points.
The Implication
If you're building anything that competes with or complements coding agents, watch these pricing moves closely. OpenAI and Anthropic are calibrating willingness-to-pay in real time, and the data says developers value reliable access to agentic tools more than most consumer AI features. For developers, the question is simple: are you getting $100/month of value from Codex versus Claude? That's the new benchmark. For companies, this is the early signal that your engineering team's tooling budget is about to look very different. Plan accordingly.
Sources: The Verge AI | VentureBeat | TechCrunch AI