Top AI News - July 01, 2026

Grok 4.5, based on a 1.5T parameter model, is now in private beta at SpaceX and Tesla with performance reportedly matching Claude Opus. SpaceX will release new models monthly in 2024.

OpenAI

GPT-5 Pro helped immunologist Derya Unutmaz solve a three-year immunology puzzle by identifying how deoxyglucose affects T cell specialization and correctly predicting unpublished experimental outcomes, demonstrating AI's potential to dramatically accelerate biological and cancer research while requiring continued human expertise for validation.

The Trump administration has effectively created an involuntary licensing system for frontier AI models without clear safety standards, threatening the economics of AI development. The proposed solution involves independent verification organizations auditing labs' compliance with their own safety frameworks, gaining support from Congress and industry.

PR Newswire

Quantifind raised $200M for its AI-native risk intelligence platform that helps six of the top 10 global banks detect financial crime with fewer false positives, potentially saving Tier 1 banks up to $177.9M annually. The company is developing agentic middleware to enable AI agents to autonomously operate within regulated risk workflows while maintaining compliance and human oversight.

NVIDIA Research

NVIDIA's ENPIRE enables robots to autonomously improve their own policies through real-world learning, reducing the need for human intervention and advancing practical robotics deployment capabilities.