OpenAI launches GPT-5.6 and ChatGPT Work
OpenAI’s latest model family arrives alongside a workplace agent built for multi-step knowledge work.
OpenAI is pairing a new model generation with a more explicit push into everyday workplace workflows.
What happened
OpenAI launched the GPT-5.6 family with three variants — Sol, Terra and Luna — and introduced ChatGPT Work, a workplace-focused agent designed for multi-step tasks involving documents, spreadsheets and presentations. The models are being rolled out across ChatGPT, Codex and the OpenAI API, with the lineup positioned around coding, enterprise work and scientific use.
Why it matters
The release is less about a single benchmark jump and more about product segmentation. OpenAI is increasingly matching different model tiers to different workloads and cost profiles, while ChatGPT Work pushes the company further into the software layer where AI systems execute tasks rather than simply answer prompts.
The bigger picture
The frontier-model market is becoming a competition over complete workflows. Model quality still matters, but distribution, agent capabilities, enterprise integration and cost per completed task are becoming just as important.