GPT-5.4 Pro Goes Live: OpenAI's Reasoning Beast Now in ChatGPT & API!
The new reasoning-focused model arrives as OpenAI faces internal debate over its Pentagon partnership.
OpenAI has officially released GPT-5.4 Pro, a new model touted as a 'reasoning beast' now accessible through ChatGPT and its developer API. This launch represents a significant step in the company's push toward more advanced, logical AI systems, though it arrives during a period of intense scrutiny for the company.
Simultaneously, OpenAI is grappling with substantial internal and external criticism following reports of an expanded partnership with the U.S. Department of Defense. CEO Sam Altman acknowledged that communication around the Pentagon agreement was 'sloppy,' and The Guardian reported that a senior robotics executive resigned over the controversy. This incident underscores the growing tension between frontier AI companies' commercial ambitions and the ethical boundaries of military applications.
The week's news also highlighted a broader industry shift, as U.S. federal agencies reportedly began phasing out Anthropic's Claude models due to disagreements over military use restrictions. This move could funnel more government contracts toward OpenAI and Google, turning AI vendor alignment into a geopolitical issue. Meanwhile, the industry continues to advance on other fronts, with KDDI and Avita partnering on humanoid service robots and AI PCs taking center stage at Mobile World Congress, signaling a move toward embodied and edge-based AI.
- OpenAI launched GPT-5.4 Pro, a new reasoning-optimized model now live in ChatGPT and API.
- A senior OpenAI robotics executive resigned amid controversy over an expanded Pentagon deal, which CEO Sam Altman called 'sloppy'.
- U.S. agencies are reportedly phasing out Anthropic's Claude models over military use disputes, potentially benefiting OpenAI.
Why It Matters
The launch amplifies the conflict between rapid AI capability growth and the urgent need for ethical governance in high-stakes applications.