ADNOC and G42 signed a multi-year agreement with OpenAI on November 2 that makes the Emirati state oil company the first in the energy sector to deploy OpenAI Frontier, running GPT models on inference infrastructure inside the UAE. It was one of four agreements signed as Crown Prince Sheikh Khaled met Sam Altman, alongside a Mubadala deal on AI-driven investment analysis and a five-year push to spread frontier AI across government, enterprise and education, plus a machine-readable government initiative.
SemiAnalysis released GDPx, an agentic benchmark spanning 11 high-value domains including chip design, cybersecurity, biology, finance, law, quant trading, software engineering and recursive self-improvement. Anthropic's Claude Opus 5.5 led the debut leaderboard at 12.0%, ahead of Gemini 4 Carbon at 10.9% and GPT-6 Astra at 9.6%, while finance and legal tasks left nearly every model at zero; the firm noted safety classifiers also suppressed some scores.
Meituan launched LongCat-2.5 on November 2, a multimodal model with 1.6 trillion total parameters and roughly 48 billion activated per token, designed for tasks where AI agents must keep working for a long time. It was trained on long task sequences and large-scale reinforcement learning to handle complex codebases and complete software projects, processes images and text in one token framework, and is available through Meituan's chat service and API, with GitHub and Hugging Face releases to follow.
OpenAI is preparing a public-relations push after Tuesday's US midterm elections and is considering rebranding its data centers as "AI compute campuses," Axios reported. Chief global affairs officer Chris Lehane also floated AI "speed limits," efforts to show AI as empowering in daily life and a US role in setting global standards, calling the backlash over data centers "good and healthy" while arguing communities should hold companies accountable.
Chinese embodied-AI companies raised 93.5 billion yuan ($13 billion) across 322 deals in the first half of 2026, five times the amount a year earlier, according to research firm ITjuzi. After years of showing humanoids that run, dance and somersault, robotics firms are directing capital toward software that lets machines perceive their surroundings, learn unfamiliar tasks and adapt without hand-written code for every action.