OpenAI will publicly launch its GPT-5.6 model family (Sol, Terra, Luna) on Thursday July 9, after delaying the release in June at the Trump administration's request over national security concerns. GPT-5.6 Sol scored 88.8% on TerminalBench 2.1, outpacing Anthropic's Claude Opus 4.8 at 78.9% in coding benchmarks.
Chinese authorities have held talks with Alibaba, ByteDance, and Z.ai over the past month about restricting foreign access to their most advanced AI models, including open-weight releases not yet published. A proposed tiered system would require basic open-source models to file simple registrations, while the most sensitive frontier systems could be barred from public release or restricted to domestic use.
Reuters reports that Chinese AI startup DeepSeek is designing its own inference chip, reportedly to be fabbed at SMIC, to reduce dependence on Nvidia and Huawei. Chinese firms plan to route 46% of their AI-accelerator budgets to domestic chips over the next year, up from 30% today, as US export controls limit access to advanced hardware.
Microsoft has begun deploying its own MAI-Thinking-1 model across productivity apps like Excel and Outlook, replacing OpenAI and Anthropic models to cut costs. In blind tests across 1,276 tasks, evaluators preferred MAI-Thinking-1 over Claude Sonnet 4.6; it scores 52.8% on SWE-Bench, matching Claude Opus 4.6 on coding.
Meta's Superintelligence Labs, led by Chief AI Officer Alexandr Wang, released its first image generation model, Muse Image, rolling out inside the Meta AI chatbot, Instagram, and WhatsApp. The model can build images from scratch, edit photos, blend references, and pull real-time web context. Meta says Muse Video is already in development.
AI chip startup SambaNova raised $1 billion at an $11 billion valuation, with JPMorgan Chase selecting it as an inference infrastructure partner. Separately, Samsung's quarterly profit surged 19-fold on runaway AI memory chip demand, though shares fell 10% as investors had expected even stronger results.
Governor JB Pritzker signed the AI Safety Measures Act, making Illinois the first state to require independent safety audits of frontier AI developers and mandate reporting of critical safety incidents. With similar bills in California and New York, the three states cover roughly 40% of the US AI market, creating a de facto national standard.