SpaceXAI — the AI division born from Elon Musk's decision to fold xAI into SpaceX — has released Grok 4.5, a 1.5-trillion-parameter model built jointly with the AI coding startup Cursor and positioned to compete directly with Anthropic's Claude Opus and OpenAI's GPT-5.5.

The launch, reported by The Information and confirmed by Musk on July 8, 2026, represents the first major model release under the SpaceXAI banner. The company adopted the new name just one day earlier, on July 7, retiring the xAI brand it had used since 2023. For the latest developments across the AI industry, readers can follow our breaking AI news coverage.

A Foundation Built on V9 and Cursor Data

Grok 4.5 runs on a new V9 foundation model with approximately 1.5 trillion parameters, making it the largest system the company has shipped. According to Musk, the model incorporates data from Cursor as supplemental training — a significant detail given that SpaceX is in the process of acquiring the coding startup for $60 billion.

Musk first teased the model on June 28, posting on X: "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 & Tesla. Early evals show performance close to, perhaps exceeding Opus."

Cursor founder Michael Truell had separately signaled the collaboration, announcing last month that Cursor was developing a model intended to compete with frontier systems from Anthropic and OpenAI. At the time, he described it as the "next phase of the company" without explicitly disclosing the SpaceXAI partnership.

Performance Claims: Opus-Class at Lower Cost

According to reporting from The Information, the model is being compared internally to Claude Opus 4.8 — Anthropic's latest Opus-tier release — as well as OpenAI's GPT-5.5. Musk has stated that Grok 4.5 is both faster and cheaper than competing Opus-class models, positioning it as a cost-effective alternative for developers and enterprises.

However, these performance claims rest on internal evaluations conducted at SpaceX and Tesla, where the model reportedly ran in private beta for approximately two weeks before release. Independent benchmark numbers were not available at launch.

The stakes are considerable. Claude Opus sits just below Anthropic's most powerful — and most controversial — model, Fable, which was subject to a government review earlier this year. GPT-5.6, OpenAI's expected next release, has not yet shipped publicly.

Access and Pricing

References to Grok 4.5 surfaced across SpaceXAI products in the days leading up to launch. A feature flag discovered in the Grok web client displayed alternate upgrade-page copy naming the model, while the Grok Build CLI — the company's terminal-based coding agent — prompted users to move to the SuperGrok Heavy plan before accessing it.

The SuperGrok Heavy plan is currently offered at a promotional rate of $99 per month, though it remains unclear whether that price will persist after the launch window. Access appears to follow the gated rollout pattern SpaceXAI has used for prior releases.

Strategic Context: The SpaceXAI Rebrand

The release carries outsized strategic weight for SpaceXAI. The rebranding from xAI to SpaceXAI consolidates Musk's AI ambitions under the SpaceX corporate umbrella, giving the division access to the aerospace company's engineering talent and capital base. Musk has already directed top Starship and Starlink engineers to work on Grok, according to a June report from Business Insider.

The $60 billion Cursor acquisition — SpaceXAI's first major deal under the new name — is still pending. Grok 4.5's launch ahead of that deal closing represents a calculated bet: demonstrating that the combined entity can deliver a frontier-competitive model before regulators and shareholders sign off on the merger.

For SpaceXAI, the model is ultimately a test of whether the compute, data, and capital gained through the SpaceX merger and Cursor acquisition can narrow the gap with established leaders like Anthropic and OpenAI. Musk has acknowledged that earlier release targets — originally set for late spring — slipped, and whether Grok 4.5 can sustain its claimed performance edge against upcoming releases from competitors remains the open question.

Stay Ahead of AI

The AI model landscape is moving faster than ever. For comprehensive coverage of new releases, benchmark battles, and industry strategy, visit AI Buzz Wire.

Read more AI news →