xAI Grok 4.5 enters private beta at Tesla and SpaceX — Musk claims near-Opus performance
Elon Musk announced on X that xAI's Grok 4.5 — a 1.5-trillion-parameter model built on the V9 foundation and supplementally trained on Cursor coding data — has entered private beta at both Tesla and SpaceX. Early internal benchmarks, per Musk, put its performance close to or above Claude Opus. It is xAI's first publicly confirmed deployment of its own AI model inside Tesla's engineering environment.
Grok 4.5 is built on xAI's V9 foundation model with 1.5 trillion parameters — a 50% step up from Grok 4.4's 1 trillion. Training wrapped on May 26, 2026, with a key supplemental addition: coding data from Cursor, a popular AI-assisted development environment. That integration specifically sharpens the model for software engineering and technical reasoning, making it directly relevant to the workflows at both Tesla and SpaceX.
The most significant fact for Tesla shareholders is where the first deployment is happening. By entering private beta specifically inside Tesla and SpaceX, Grok 4.5 marks xAI's first publicly confirmed AI model integration into Tesla's engineering environment. This is currently productivity-tool-level deployment for engineers — not a direct FSD or vehicle feature integration — but it establishes the precedent and direction: xAI capabilities converging with Tesla's software stack.
Musk also signaled an accelerated roadmap: xAI intends to release new AI models 'completely trained from scratch' through SpaceX every month for the rest of 2026. The next model iteration is already in reinforcement learning training, per the tweet — an unusually fast cadence for models at this scale and parameter count.
On performance: Musk's claim that early evals put Grok 4.5 at or above Claude Opus would place it in the top tier of available large language models, alongside GPT-4o and Gemini Ultra. Internal benchmark claims are not the same as independent third-party evaluations, but the deployment at Tesla and SpaceX — with real engineering workflows as the test bed — gives these evals more grounding than a purely synthetic benchmark. The story has been picked up by Seeking Alpha, Basenor, and Crypto Briefing, confirming multiple cross-sources for the announcement.
Summaries are prepared by the Tesla Briefing editorial team and may not capture every nuance of the original reporting. You are solely responsible for your own investment decisions.