OpenAI and SpaceXAI Launch New AI Models Amid Government Scrutiny

OpenAI and SpaceXAI both unveiled new AI models this week, with OpenAI's GPT-5.6 family and SpaceXAI's Grok 4.5 entering public availability amid heightened government scrutiny over national security risks. Both launches follow months of evaluation by the Trump administration, which imposed temporary restrictions on advanced AI models.

OpenAI's GPT-5.6 Family

OpenAI confirmed GPT-5.6 Sol, Terra, and Luna will be available to the public starting Thursday, July 9. Sol is the most powerful model yet, with improvements in programming, biology, and cybersecurity, while Terra and Luna offer intermediate and low-cost options Source: portafolioco. The company emphasized that GPT-5.6 Sol incorporates its "most robust safety set" and underwent weeks of vulnerability testing Source: portafolioco. Axios reported that the Trump administration gave the green light after technical reviews Source: elcolombiano.

SpaceXAI's Grok 4.5

SpaceXAI (formerly xAI) introduced Grok 4.5, developed in collaboration with Cursor, a coding startup SpaceX recently agreed to acquire for $60 billion Source: larepublicaco. The model is optimized for programming, legal, and financial tasks, with 80 tokens per second inference speed and a cost of $2 per million input tokens Source: infobae. Elon Musk called it "more competitive" than Anthropic's Opus 4.7 in real-world applications Source: infobae.

Government Oversight

The launches come after a period of tight government control. In June, President Donald Trump signed an executive order establishing a voluntary framework for pre-release government review of "frontier models" Source: euronewses. OpenAI initially restricted GPT-5.6 to a small group of US partners at Washington's request Source: elespanol. Similarly, Anthropic's Mythos models remain limited Source: elespanol. Both OpenAI and SpaceXAI now have clearance for broader release, though Cursor noted measures to block cybercriminals Source: larepublicaco.