SpaceXAI’s Grok 4.5 launches to the public tomorrow, July 9, 2026, as an Opus-class coding model built on a 1.5 trillion parameter V9 foundation, trained with data from the AI code editor Cursor. The release marks the company’s first major model launch since going public and acquiring Cursor earlier this year.
Elon Musk described Grok 4.5 as “an Opus-class model, but faster, more token-efficient and lower cost,” according to a post on X. The model will be available immediately in Grok Build, across all Cursor plans, and via the SpaceXAI console, though it won’t launch in the EU tomorrow.
The model’s training combined xAI’s V9 foundation — which the company says is roughly three times larger than its previous v8-small architecture — with supplemental training using Cursor’s coding data, a partnership that gives Grok 4.5 specialized strength for engineering and knowledge work. Early internal evaluations reportedly put it close to or ahead of Anthropic’s Claude Opus on several benchmarks.
Pricing is a major competitive angle. Grok 4.5 costs $2 per million input tokens and $6 per million output tokens, according to Axios, compared to Claude Opus 4.8’s $5 input and $25 output pricing. That cost advantage could appeal to developers building agentic systems and automation workflows, even if Grok 4.5 does not yet match the performance of OpenAI’s latest models.
The launch comes as SpaceXAI competes for market share in a crowded frontier AI landscape. Anthropic’s Claude Opus family has dominated coding benchmarks; Claude Opus 4.8 currently leads on the SWE-bench Verified benchmark, a measure of real-world software engineering tasks. Grok 4.5’s positioning as a faster, cheaper alternative may help xAI capture cost-conscious teams, even if the model trails the absolute best performers from Anthropic and OpenAI on raw capability.
SpaceXAI has also said it plans to release a new foundation model every month through the end of 2026, signaling an aggressive roadmap to narrow the capability gap with its rivals.
Sources
- Axios — confirmed Grok 4.5 launch date, Opus-class positioning, V9 foundation, 1.5 trillion parameters, Cursor training data, pricing ($2/$6 vs. Opus 4.8 at $5/$25), and availability across Grok Build, Cursor, and SpaceXAI console.
- Mashable — confirmed Grok 4.5 public release on July 9, Opus-class model, faster and lower-cost positioning.
- Threads — confirmed Musk’s statement that Grok 4.5 is an Opus-class model but faster, more token-efficient, and lower cost; confirmed launch date.
- Digital Applied — confirmed V9 foundation model at 1.5 trillion parameters, Cursor data in supplemental training, and private beta at SpaceX and Tesla.
- LM Council — confirmed Claude Opus 4.7 leads coding benchmarks at 83.5% on recent evaluations.
