Editor's Summary

This was a week defined by frontier model rollouts and by AI's growing entanglement with government scrutiny, as OpenAI and xAI pushed new models into production while regulators and enterprise partners shaped how they get used.

OpenAI had the busiest week of any vendor. It began rolling out its top-tier model globally after a limited preview, and separately confirmed that GPT-5.6 is now the "preferred model" powering Microsoft Copilot 365, a signal worth watching given ongoing chatter about strain in the Microsoft-OpenAI relationship. Sam Altman also said the company adjusted its models following discussions with US officials, underscoring how directly Washington is now shaping frontier model behavior rather than just its deployment.

xAI matched that pace with Grok 4.5, introduced this week and immediately positioned for enterprise use: SpaceX and Cursor are backing a push into finance and legal applications, a notable pivot from Grok's consumer-chatbot origins toward regulated, high-stakes verticals.

Anthropic's week was more infrastructure- and product-focused. The Government of Alberta disclosed it is using Claude to find and fix cybersecurity vulnerabilities across its government systems, a concrete public-sector deployment rather than a pilot announcement. Anthropic also continued building out Claude Cowork, its coding-agent product for broader office work, with plans to expand it to mobile and web — part of a wider trend of coding agents moving beyond development teams into general knowledge work.

Taken together, the week's secondary stories reinforce a theme: AI vendors are no longer just shipping models, they're embedding themselves in government workflows (Alberta's cybersecurity use of Claude, Altman's acknowledgment of US official input) and into everyday enterprise software (Copilot 365's shift to GPT-5.6, Grok 4.5's finance and legal push).

To watch next week: whether the Microsoft-OpenAI relationship tension referenced alongside the GPT-5.6 Copilot news develops into something more concrete, and how quickly Grok 4.5's finance and legal positioning translates into actual enterprise adoption.

Friday, July 10, 2026
Apple Sues OpenAI for Trade Secret Theft Over AI Hardware Designs
Thursday, July 9, 2026
OpenAI launches its new family of models with GPT-5.6
OpenAI says GPT 5.6 is the 'preferred model' for Microsoft Copilot 365 amid breakup chatter
Altman Says OpenAI Adjusted AI Models After Discussions With US Officials
Introducing a way to reflect on how you use Claude
Meta's new AI chips will begin production in September
Wednesday, July 8, 2026
OpenAI to Roll Out Top AI Model Globally After Limited Preview
SpaceXAI, Cursor Launch Grok 4.5 AI Model for Finance, Legal Applications
Introducing Grok 4.5
xAI launches Grok 4.5, its first release since acquiring Cursor
Automating cross-repo documentation with GitHub Agentic Workflows
How GitHub Copilot enables zero DNS configuration for GitHub Pages
Tuesday, July 7, 2026
The coding agent wars are spilling into the rest of the office: Claude Cowork
Anthropic expanding Claude Cowork to mobile and web, details here
Microsoft Replaces OpenAI, Anthropic With Own AI in Some Apps
Introducing Muse Image and Muse Video
GitHub Copilot app available to all
Kimi K2.7 now available for Copilot Business and Enterprise
Monday, July 6, 2026
Government of Alberta uses Claude to find and fix cybersecurity vulnerabilities across government systems
Anthropic says Claude has carved out its own space to ponder
The Making of Claude Code
CFOs and the new economics of AI