OpenAI says its new GPT-5.5 model is more efficient and better at coding

OpenAI has announced GPT-5.5, its latest large language model, calling it 'the smartest and most intuitive to use model yet, and the next step toward a new way of getting work done on a computer.' Released just one month after GPT-5.4, the new model excels at writing and debugging code, conducting online research, creating spreadsheets and documents, and performing work across multiple tools. According to OpenAI, users can now 'give GPT-5.5 a messy, multi-part task and trust it to plan, use tools, check its work, navigate through ambiguity, and keep going.' The company also noted that GPT-5.5 uses 'significantly fewer' tokens to complete tasks in Codex, and features its 'strongest set of safeguards to date.'

GPT-5.5 will roll out starting Thursday to Plus, Pro, Business, and Enterprise ChatGPT tiers, as well as Codex, with a GPT-5.5 Pro version coming to Pro, Business, and Enterprise users. The release is part of an escalating AI arms race with Anthropic, which recently launched Claude Opus 4.7 and announced Mythos Preview, a non-public model focused on cybersecurity. OpenAI countered with GPT-5.4-Cyber, trained to flag vulnerabilities. Both companies are competing for enterprise and coding market share, with OpenAI shifting focus from 'side quests' to larger revenue streams. The launch also comes days before the Elon Musk vs. OpenAI trial, set to begin Monday in Oakland, California.