On X, the current claim is that Grok 4.5 posted a strong showing on Zapier's AutomationBench while also costing less to run than rival models. Elon Musk described it as Grok 4.5 being "#1 in some respects," and posts on X are echoing benchmark claims about leaderboard performance and lower per-task cost. So far, though, the public sourcing in view is still social posts pointing back to the benchmark image and Musk's post, not an independently reported write-up or full benchmark documentation.