OpenAI's GPT-5.5 Delivers Major Performance Gains, Expands Codex's General-Purpose Computing Capabilities

OpenAI has not publicly confirmed GPT-5.6, but ChatGPT users this week have widely reported more coherent responses and deeper reasoning, along with longer wait times. This has fueled speculation that the platform is conducting A/B testing with a subset of accounts. If true, OpenAI is testing the performance and computing costs of its next-generation model, with potential implications for competition in the generative AI market.

Recent hands-on tests indicate that the suspected new model excels at complex tasks such as generating 3D games and simulating robots, with some users describing it as better than Fable 5. Market rumors suggest GPT-5.6 could be released as early as June 25. However, OpenAI has yet to disclose the scale of the testing, model parameters, pricing or a firm timetable, and the claims still await official confirmation.

Large language models are evolving from single-turn question-answering tools into agentic systems that can independently plan, use tools and complete multistep tasks. OpenAI is positioning GPT-5.5 as the primary workhorse for ChatGPT and Codex, targeting software development, data analysis, document creation and scientific research. The move shows that competition among AI models has expanded into enterprise workflows and research output.

OpenAI released GPT-5.5 on April 23, 2026, and made it available through the API on April 24. Official specifications list a 1 million-token context window for the API and 400,000 tokens for Codex, not 12 million. The model topped the Artificial Analysis Intelligence Index and scored 82.7% on Terminal-Bench 2.0. Pricing is $5 per million input tokens and $30 per million output tokens.

OpenAI and Anthropic are competing for the AI software development tools market, with Codex going head-to-head with Claude Code. OpenAI President Greg Brockman has instructed engineering teams to make AI agents central to their daily development work. The aim is to shorten coding, testing and deployment cycles, deepen the model’s adoption in enterprise software workflows and strengthen the ChatGPT ecosystem.

OpenAI recently began a six-week product sprint that includes rapid releases of the Codex desktop app, GPT-5.4 and security-agent features. It also plans to combine ChatGPT, Codex and the Atlas browser into a desktop “super app.” The company has not disclosed the investment involved or exact release dates. The central goal of the update push is to counter Anthropic’s Claude Code.

Since launching the GPT-5 series, OpenAI has continued to develop large language models beyond conversational tools into AI agents capable of executing multistep tasks. Released on March 5, 2026, GPT-5.4 is the company’s first general-purpose model with native computer-use capabilities. It can issue mouse and keyboard commands based on screenshots and complete tasks across websites, spreadsheets and documents, lowering the integration barrier for automating professional workflows.

OpenAI made GPT-5.4 Thinking available in ChatGPT, the API and Codex on the same day. API pricing is $2.50 per million input tokens and $15 per million output tokens. The model achieved an 83.0% win-or-tie rate on the GDPval professional-work benchmark. In Tool Search testing across 250 tasks, it reduced token usage by 47% without sacrificing accuracy.