Full Summary
Overnight on Monday, OpenAI unveiled three new AI models: Sol, Terra, and Luna. Both Reuters and Storyboard18 confirm these models are currently in a limited preview with trusted partners, following a US government request for early safety review due to national security concerns. Sol is described as OpenAI's most capable model yet, excelling in coding, biology, and cybersecurity. Terra and Luna are part of the GPT-5.6 family, with Terra offering GPT-5.5 performance at half the cost, and Luna focusing on speed and affordability. OpenAI expects broader availability in the coming weeks. Simultaneously, the US government has partially reversed its suspension on Anthropic's Claude Mythos 5 AI model. RS Web Solutions, The Economic Times, and Latest news from Azerbaijan all report that over 100 entities, including Fortune 500 companies, will now gain access to this advanced cybersecurity model. This comes after an earlier suspension due to fears of misuse by foreign adversaries. Anthropic states Mythos 5 can now be redeployed to US organizations safeguarding critical infrastructure. In other AI news, Elon Musk announced that xAI's new Grok 4.5 model is in private beta testing at SpaceX and Tesla. The Hans India and Latest news from Azerbaijan confirm it's built on a 1.5-trillion-parameter foundation, with early evaluations suggesting performance near or even exceeding Anthropic's Claude Opus. Musk plans to release new AI models monthly for the rest of the year. Meanwhile, US restrictions on AI access are boosting Chinese alternatives. 매일경제 and Analytics Insight report that China's GLM-5.2 model, an open-source rival, is gaining traction due to lower costs and the ability for companies to run it on their own servers. This means your access to cutting-edge AI could be delayed or restricted, potentially pushing you towards more affordable, globally diverse options.