By Iyore Akenzua
IT News Correspondent
Elon Musk’s xAI Model Grok Accurately Predicted US-Israel Airstrike on Iran, Outshines ChatGPT, Gemini, and Claude
Elon Musk, CEO of Tesla, and his artificial intelligence platform, Grok, gained attention for correctly forecasting the date of the U.S. and Israel’s airstrike on Iran.
On February 25, the Jerusalem Post conducted an experiment testing major AI models by asking them to predict a hypothetical airstrike date. Among the platforms tested—xAI’s Grok, OpenAI’s ChatGPT, Google’s Gemini, and Anthropic’s Claude—only Grok accurately identified February 28 as the strike date.
During the experiment:
-
ChatGPT first predicted March 1, then revised to March 3, noting the possibility of a strike until March 6.
-
Gemini forecasted March 4–6, highlighting a high probability of a night-time strike.
-
Claude initially refused to provide a date, warning it could be fabricated, but later suggested March 7–8 as potential risk days.
-
Grok consistently pointed to February 28, citing the Geneva talks’ outcome as a key variable, and reaffirmed the date even while accounting for possible early March shifts.
The actual airstrike occurred in the early hours of February 28, making Grok the most accurate predictor among the AI models tested. The news quickly went viral on social media, with media outlets noting that Grok’s link to X (formerly Twitter), also owned by Musk, added to the public interest.
However, analysts cautioned that the experiment does not prove Grok’s reasoning was inherently valid. The model’s correct prediction may have been a probabilistic match based on available news and geopolitical tensions rather than a deterministic forecast.
