Someone on X asked Elon Musk how Grok 4.8 would compare with Claude Opus 5.5. His reply was blunter than the marketing. "Not as good. Opus 5.5 is a masterpiece. But Grok 5 should exceed Opus 5.5 and maybe that happens in 2 or 3 months, with continued acceleration thereafter."
Not as good. Opus 5.5 is a masterpiece. But Grok 5 should exceed Opus 5.5 and maybe that happens in 2 or 3 months, with continued acceleration thereafter.
— Elon Musk (@elonmusk) November 10, 2026
The concession reads differently next to what he said in October, when he promised that Grok 4.8 would not be a disappointment. It is now ranked second by its own author, behind a model from Anthropic.
Independent numbers back the ranking, at least for now. The Artificial Analysis Intelligence Index v5 put Claude Opus 5.5 at the top of its table last month, ahead of every open-weight model.
Grok 4.8 is still arriving. Musk has been rolling it out to Grok Bot users, and the assistant has had a noisy few weeks over usage limits and overnight fixes. Telling those users that the new model is not the best one available is an unusual marketing choice, though it is hard to argue with the honesty.
The awkward part is that Grok Bot already runs on Claude Opus 5.5 for harder work, alongside Midjourney and Suno. So the product depends on the model Musk just called a masterpiece, while the model carrying the Grok name trails it. Most users will not care which engine answers, but the admission sharpens the question of what the Grok models are for.
Two or three months is a short and unusually specific window, and it is a claim rather than a schedule. Musk has given dates before that moved, and xAI has published no benchmark, parameter count or release date for Grok 5. The only evidence for the target is the post itself.
There is also a moving-target problem. Beating Opus 5.5 in a few months means beating the Opus 5.5 that exists now, not whatever Anthropic ships by then. That is the part of the answer Musk gets right. "Continued acceleration thereafter" is a promise about a trajectory, not about a single result, and it leaves the race open in both directions.
Comments