OpenAI reworks ChatGPT models and Codex pricing
OpenAI removed several 'o4' and GPT‑4.1 variants from ChatGPT while keeping API access for developers and tying model availability to product contexts. (help.openai.com). At the same time the company expanded Codex access and introduced a $100 Pro plan as it positions coding copilots as paid infrastructure, with an updated Codex rate card published. ( )
OpenAI has stripped several older models out of ChatGPT while leaving them on sale through its application programming interface, or API, for developers. (help.openai.com) The company says GPT-4o, GPT-4.1, GPT-4.1 mini, OpenAI o4-mini, and GPT-5 Instant and Thinking were retired from ChatGPT on February 13, 2026. GPT-5.1 Instant, GPT-5.1 Thinking, and GPT-5.1 Pro followed on March 11, 2026. (help.openai.com) OpenAI says those retired models still remain available through the API, and old chats and custom GPTs are being moved onto GPT-5.3 Instant or GPT-5.4 Thinking and Pro equivalents. ChatGPT Business, Enterprise, and Edu customers kept GPT-4o inside custom GPTs until April 3, 2026. (help.openai.com) At the same time, OpenAI is narrowing which models people can pick directly inside ChatGPT and pushing more of the product toward task-based routing. In March, it rolled out GPT-5.4 mini as a fallback for GPT-5.4 Thinking instead of a model users can manually select from the picker. (help.openai.com) That leaves ChatGPT looking less like a shelf of model names and more like a bundled app with background switching. OpenAI says GPT-5.4 Thinking combines its newer reasoning, coding, and agent-style workflow work into one model for spreadsheets, documents, software tools, and web research. (help.openai.com) The pricing move lands on the coding side. OpenAI’s help center says Codex pricing changed on April 2, 2026 from per-message charges to token-based billing, the same basic meter used in APIs that counts input, cached input, and output. (help.openai.com) Under the new rate card, GPT-5.4 costs 62.50 credits per 1 million input tokens and 375 credits per 1 million output tokens, while GPT-5.3-Codex costs 43.75 input credits and 350 output credits per 1 million tokens. OpenAI says fast mode uses twice as many credits. (help.openai.com) OpenAI also added a new $100-a-month Pro tier for ChatGPT. The company says that plan gives 5 times the usage of Plus, and 10 times the Codex usage of Plus for a limited period, while the older $200 Pro tier remains in place with 20 times the usage of Plus. (help.openai.com) On OpenAI’s Codex pricing page, the company says Pro now starts at $100 a month and includes priority request processing and access to GPT-5.3-Codex-Spark, a research-preview coding model. The same page says the temporary Codex usage boost ends May 31, 2026. (developers.openai.com) OpenAI’s help center says Codex now sits inside Free, Go, Plus, Pro, Business, Edu, and Enterprise plans, with extra credits available after limits are hit. The company says average Codex spend runs about $100 to $200 per developer per month, depending on model choice, cloud usage, automations, and fast mode. (developers.openai.com)