When will any company achieve AGI?

The strongest case for Yes is OpenAI’s unusually direct public signaling. Reporting on Sam Altman’s TIME interview indicates the company expects to have an internal system by the end of 2026 that he would personally call AGI, and OpenAI’s own definition is broad enough to make a declaration possible if they believe the system clears their bar. That matters because this event does not require consensus validation or an external benchmark, only a company announcement that it has achieved AGI.

The biggest reason to stay below a majority probability is that this is still a self-authored label with major ambiguity. A company can make rapid progress without choosing to use the exact phrase AGI, especially if the model is impressive but controversial, if the claim invites regulatory or investor scrutiny, or if the system falls short in ways that make the announcement look overstated. Prediction-market commentary already reflects this hesitation, and the contract wording appears strict about requiring an explicit AGI claim rather than softer language like AGI-like or near-AGI.

Compared with the current market price of 12%, my estimate is meaningfully higher because the event is more of a corporate communication decision than a universal scientific breakthrough, and OpenAI has now publicly planted a flag on the timing. Still, 22% is well below 50% because many things must go right at once: technical progress, internal confidence, and a willingness to make a legally and reputationally loaded public statement before the deadline. The market looks a bit pessimistic, but not wildly mispriced.