streugut.ch
Archive

AI competition is moving away from frontier models alone toward control of the stack that turns models into governed execution. That shift is being carried by agent environments tied to enterprise data, workflow, security, measurement, pricing, and distribution, and by model access that is increasingly routed through managed clouds and owned work surfaces rather than offered as a free-standing release. The infrastructure race underneath remains a gigawatt-scale systems buildout around power, capacity, and financing. What is particular now is that this control boundary is becoming more explicit at the point of use: the same platforms are widening model choice while tightening admin permission, deployment paths, runtime security, and cost visibility inside their own systems.

The move

Microsoft puts OpenAI’s GPT-5.6 into Microsoft 365 Copilot as OpenAI launches ChatGPT Work

The clearest new move came on July 9, when Microsoft announced that OpenAI’s GPT-5.6 model family is now available across Microsoft 365 Copilot apps and is the preferred model for Word, Excel, PowerPoint, Chat and Cowork. That changes the default model many Microsoft 365 Copilot users meet inside Microsoft’s own products. The same day, Microsoft said OpenAI’s GPT-5.6 Sol, Terra and Luna models are generally available in Microsoft Foundry, alongside hosted production agents that can be deployed into Microsoft 365 Copilot and Teams.

OpenAI made the other big move on July 9: it announced the general availability of the GPT-5.6 Sol, Terra and Luna family, and it launched ChatGPT Work, a new agent inside ChatGPT for enterprise plans and the desktop app. Microsoft also updated its Microsoft Online Services subprocessor list on July 9 to add OpenAI for customers that explicitly enable OpenAI-operated models such as GPT-5.6 within Microsoft 365 Copilot.

Among the main competitors, Anthropic announced on July 9 a partnership with UST under which UST will deploy Claude into engineering environments and train 20,000 engineers, architects and consultants globally. Google Cloud announced on July 9 that Cloud Run sandboxes are in public preview as a runtime for untrusted code and agent workloads.

Sources