Particle.news
Download on the App Store

OpenAI Launches GPT-5.6 and ChatGPT Work

The company is shifting ChatGPT from a conversational tool to autonomous, multi-step workplace agents after a U.S. government review that gated initial access.

Overview

  • OpenAI announced the GPT-5.6 model family and ChatGPT Work on July 9, 2026, rolling the models into ChatGPT, Codex and the API with tiered availability that starts with paid plans.
  • ChatGPT Work is an agentic assistant that can run scheduled, multi-step tasks across desktop, web and mobile, access connected apps and files, and produce finished deliverables such as documents, spreadsheets, presentations and sites.
  • GPT-5.6 ships in three variants — Sol (flagship), Terra (balanced) and Luna (fast/cost-efficient) — and adds configurable reasoning modes including an Ultra/Max option that coordinates multiple agents for harder problems.
  • OpenAI published per-token pricing for the series and highlighted benchmark and efficiency claims versus Anthropic, saying Sol is faster and uses fewer tokens on many coding and agent benchmarks.
  • The staged rollout followed weeks of U.S. government consultation, and coverage warns that enterprises will need new workflows, governance and data-residency plans to safely operationalize agentic automation while IT services and compliance teams adapt.