OpenAI Launches GPT-5.6 Sol Under US Government Access Control

OpenAI releases GPT-5.6 Sol, Terra, Luna under government-gated preview OpenAI announced three new models — flagship Sol, mid-range Terra, and fast/cheap Luna — but is initially limiting access to a small group of government-approved partners at the Trump administration's request. The models are classified as 'High' capability in cybersecurity and biological/chemical risk, though none reached the critical threshold for autonomous cyberattacks. OpenAI stated it does not believe this government access process should become the long-term default. 123456

OpenAI and Broadcom unveil Jalapeño inference chip for LLMs OpenAI and Broadcom announced Jalapeño, OpenAI's first custom inference accelerator designed from the ground up for LLM workloads. The chip went from design to production in nine months, accelerated by OpenAI's own models, and early testing shows substantially better performance per watt than current state-of-the-art. The chip is part of a multi-generation compute platform to be deployed at gigawatt scale. 7

OpenAI leaning toward delaying IPO until 2027, NYT reports According to the New York Times, OpenAI is considering postponing its initial public offering until 2027 due to tech stock volatility and concerns about retail investor enthusiasm following SpaceX's record IPO. Bankers advising the company cautioned that market conditions could dampen demand, though OpenAI has filed confidentially with the SEC and has not made a final decision on timing. 8

Anthropic publishes Economic Index report on Claude usage cadences Anthropic released its June 2026 Economic Index report, analyzing how Claude usage patterns mirror economic rhythms — work queries dip on weekends but less in high-paid occupations, tax requests surge near filing deadlines, and sleep advice peaks around 5 a.m. The report also introduces a new survey linking usage patterns to worker expectations about AI's impact on pay, job security, and meaning. 9

Two more Google DeepMind researchers reportedly leaving for Anthropic Bloomberg reports that Google DeepMind researchers Jonas Adler and Alexander Pritzel are expected to join Anthropic, following recent high-profile departures including Nobel Prize winner John Jumper. Adler worked on AI coding initiatives and Pritzel specialized in training large language models; both were key contributors to Google's Gemini AI systems. 10