Elon Musk announced on June 28 that Grok 4.5 has entered private beta at SpaceX and Tesla, running on a 1.5 trillion parameter V9 foundation model — a 50% scale increase from Grok 4.4 in roughly one month.
"Grok 4.5, based on our 1.5T V9 foundation model, with Cursor data added in supplemental training, is now in private beta at SpaceX and Tesla," Musk posted on X. "Early evals show performance close to, perhaps exceeding Opus."
The Cursor training data comes from SpaceX's $60 billion acquisition of Anysphere in June 2026. Cursor IDE session data was used specifically to sharpen coding and technical reasoning. No independent benchmarks exist yet; early tester Mehul Mohan described the experience as "similar to Opus."
xAI is using SpaceX and Tesla as production evaluation environments — aerospace engineering and vehicle software development provide real-world tasks harder than any benchmark. The model is roughly three times larger than the production Grok on X.
Analysis
Musk's claim of "exceeding Opus" is unverified, but the deployment strategy is shrewd. Testing inside SpaceX and Tesla gives xAI closed-environment feedback without exposing an unfinished model to public benchmarks. The real question is whether Grok 4.5 ships to developers or stays locked inside Musk's companies.
