Grok 4.7 faces GPT-6 Astra and Fable 5, with engineering as its edge
Ahead of Grok 4.7's release, a user argues its only real advantage is SpaceX engineering and rocket test data, while Grok 4.6 already trails Fable 5 and GPT-6 Astra on common benchmarks.
xai will soon drop grok 4.7 and its f*cking insane can it beat fable 5 and gpt-6 astra a bigger ai usually knows more and can do more things but bigger doesn't always mean better nobody outside xai has tested grok 4.7 yet grok 4.6, the version already out, is behind both fable 5 and gpt-6 astra on the tests people use to compare these models grok 4.7's one real edge is - spacex's own engineering data - rocket test logs and stuff no other ai has touched that could make it unusually good at engineering and robotics tasks specifically so can it win maybe, on engineering beating fable 5 and astra everywhere else is a bigger claim