Curated AI news from the last 24 hours — models, agents, enterprise, and developer tools.
OpenAI publicly released GPT-5.6 in three tiers - Sol, Terra, and Luna - priced from $1 to $5 per million input tokens depending on tier, available through ChatGPT, Codex, and the API. The launch also introduced ChatGPT Work, an enterprise agent tool for drafting documents and coordinating team tasks, and GPT-5.6 is billed as OpenAI's strongest cybersecurity model yet.
techcrunch.com ↗xAI released Grok 4.5, pitched primarily as a coding and agentic-work model rather than a consumer chatbot. It is priced at $2 per million input tokens and $6 per million output tokens, and is available in Grok Build, in Cursor on all plans, and via the xAI console, though not yet in the EU.
axios.com ↗Anthropic researchers identified an internal workspace, which they nicknamed "J-Space," that Claude appears to use to hold and manipulate ideas before putting them into words. The company says the structure shows intriguing similarities to how humans consciously access thoughts, though it stops short of claiming the finding proves machine consciousness.
axios.com ↗Microsoft's release notes for Microsoft 365 Copilot confirm that Anthropic's Claude is now selectable as a model option in Copilot Chat for complex analysis and document work. The same update rolled out watermarks for AI-generated content in Copilot across Android, Windows, iOS, Mac, and Web.
learn.microsoft.com ↗