OpenAI launches $100 Pro tier
OpenAI rolled out a new $100/month ChatGPT Pro tier aimed at developers who hit coding-assistant limits, signalling these tools are moving from novelty to core dev infrastructure. The plan expands access to Codex-style coding features and gives roughly five times the coding-agent usage of the cheaper Plus plan, positioning it against Anthropic’s developer offerings. The move also comes with broader monetisation signalling — OpenAI is publicly projecting large ad revenues down the road — and the company flagged an invite-only cyber programme tied to more capable defensive tooling it won’t broadly release. (cnbc.com) (thenewstack.io) (axios.com) (gizmodo.com)
OpenAI just filled in the awkward gap between its $20 ChatGPT Plus plan and its $200 Pro plan with a new $100 option aimed at people who use its coding tools hard enough to hit limits. CNBC reported the new tier on April 9, and OpenAI’s own pitch is that it is for heavier daily Codex use rather than casual chat. (cnbc.com) (engadget.com) The immediate target is software developers who treat an artificial intelligence coding agent like a junior teammate that writes, edits, and checks code for hours at a time. OpenAI says the new $100 plan gives about five times the Codex usage of the $20 Plus plan, which is the difference between a quick helper and a tool you can lean on all workday. (techcrunch.com) (venturebeat.com) This was a hole in OpenAI’s pricing for months. The New Stack reported that Codex had passed 3 million active users, grown fivefold in three months, and left many individual developers stuck between a cheap plan that ran out fast and a $200 plan priced more like business software. (thenewstack.io) The other company sitting in the background is Anthropic. CNBC said Anthropic already sells $100 and $200 Claude tiers with higher Claude Code limits than its standard Pro plan, so OpenAI’s new price is less a random experiment than a direct answer to a rival that got there first. (cnbc.com) (techcrunch.com) That tells you where these products are heading. A year ago, paying triple digits every month for a chatbot sounded niche; in April 2026, both OpenAI and Anthropic are carving out price bands for people who use coding agents the way designers use Adobe Photoshop or teams use GitHub. (cnbc.com) (thenewstack.io) OpenAI is also signaling that subscriptions are not the whole business model. Axios reported that the company told investors it expects about $2.5 billion in advertising revenue in 2026 and as much as $100 billion by 2030, which would make ads a core business line rather than a side experiment. (axios.com) (money.usnews.com) Those ad projections rest on a huge audience assumption. Axios said the investor presentation assumes OpenAI products reach 2.75 billion weekly users by 2030, so a $100 coding plan for power users and a future ad machine for mass users fit together as two halves of the same strategy. (axios.com) At the same time, OpenAI is drawing a line around some tools it does not want to release broadly. Its February 5 post introduced “Trusted Access for Cyber,” an invite-only program for security researchers and defenders that offers more cyber-capable or more permissive models under identity and trust checks, plus $10 million in application programming interface credits. (openai.com) Gizmodo reported this week that OpenAI is still framing that cyber effort as something for a select group rather than the public, which shows the split inside the company’s product lineup: wider access and lower friction for mainstream coding, tighter gates for models that could be more easily misused. (gizmodo.com) (openai.com) So the $100 plan is not just a price change. It is OpenAI saying there is now a real middle class of users who will pay serious money every month for coding agents, while the company builds a second future business around ads and keeps its most sensitive cyber tools behind the velvet rope. (cnbc.com) (axios.com) (openai.com)