Recent releases such as GPT-6 Astra in September 2026 prompted OpenAI executives including Greg Brockman to describe an emerging AGI era based on gains in agentic computer use, coding, and cybersecurity benchmarks, yet no formal company announcement of achieved AGI has followed. Sam Altman stated in August 2026 that OpenAI was “not quite yet” at AGI while targeting an internal system by year-end and an autonomous AI researcher only by March 2028. Independent evaluations, including ARC-AGI-3 scores well below human levels, continue to show gaps in general reasoning. Traders price the 89% “No” outcome on these cautious statements, the shift toward robotics demos planned for 2027, and the absence of the broad, verified capabilities typically required for resolution.