SpaceXAI launches Grok 4.5, Musk says new AI model rivals Claude Opus at a fraction of the cost

SpaceXAI on Wednesday introduced Grok 4.5, its newest flagship model, with Elon Musk claiming it delivers Claude Opus-class performance at a much lower price. The release marks the company’s latest push into enterprise AI and software development after its $60 billion acquisition of Anysphere, the startup behind the popular AI coding assistant Cursor.

The launch comes one day before OpenAI is expected to unveil GPT-5.6, setting up another high-profile showdown in the race for advanced AI models.

SpaceXAI says Grok 4.5 is its most intelligent AI model to date, built with a strong focus on coding and agentic workloads. The company said the model was trained across tens of thousands of Nvidia GB300 graphics processing units using carefully filtered and deduplicated datasets that were scored for quality before training.

The launch comes as competition among frontier AI labs intensifies. Model makers are racing to improve reasoning, software engineering, and autonomous task execution without driving up inference costs for developers and enterprise customers.

Cursor, one of the most widely used AI coding assistants, confirmed it has already integrated the new model.

“We’ve partnered with SpaceXAI to train Grok 4.5,” said Cursor, a popular AI coding agent.

The partnership builds on SpaceX’s acquisition of Anysphere last month in an all-stock transaction valued at $60 billion. The deal gave SpaceX direct ownership of one of the fastest-growing AI developer platforms and strengthened its position in the enterprise software market.

Musk says Grok has caught up with Anthropic

The release follows a prediction Musk made in April, when he outlined an aggressive roadmap for Grok’s development.

“It will take until May to be close to Opus 4.6 and June to match and maybe exceed. Short time by normal standards, but long time in the AI arena,” Musk wrote in an X post on April 11, 2026.

After Wednesday’s launch, Musk argued that Grok had reached that milestone.

“Our internal assessment is that Grok 4.5 is roughly comparable to Opus 4.7, but much faster. The combination of capability, faster speed and lower cost is what makes it competitive. We are closing the loop on real-world usefulness, not benchmarks. Hardcore engineers at Tesla & SpaceX find Grok 4.5 genuinely useful, which is what actually matters,” Musk wrote.

The comments reflect a broader shift across the AI industry. Labs are placing greater emphasis on production performance and developer adoption rather than on benchmark scores alone. Enterprise customers increasingly care about latency, cost, and reliability as AI systems move into day-to-day business workflows.

Grok 4.5 is priced to compete

SpaceXAI is making Grok 4.5 available immediately through Grok Build, its AI coding agent integrated with Cursor, and its developer console via API access. The company expects availability across the European Union in mid-July, Reuters reported.

Pricing is one of Grok 4.5’s biggest selling points.

The model costs $2 per million input tokens and $6 per million output tokens.

Musk described the pricing strategy in a post on X.

“It is an Opus-class model, but faster, more token-efficient, and lower cost.”

For comparison, Anthropic’s Claude Opus 4.8 costs $5 per million input tokens and $25 per million output tokens. OpenAI’s GPT-5.6 Luna is priced at $1 per million input tokens and $6 per million output tokens.

Input tokens are the text, code, or other data fed into an AI model. Output tokens are the responses the model generates.

A new chapter after xAI

The launch marks another milestone in integrating Musk’s AI ambitions into SpaceXAI. Musk’s startup xAI was acquired by SpaceX in February, ending its brief run as an independent company. Musk later said xAI would no longer exist as a separate organization and would instead operate as part of SpaceXAI.

Grok 4.5 now arrives as one of the company’s flagship AI offerings, with coding, software engineering, and autonomous agents at the center of its strategy.

The timing is unlikely to be accidental. OpenAI is expected to publicly launch GPT-5.6 on Thursday, after delaying its release last month in response to U.S. government requests related to national security concerns about the potential misuse of advanced AI technology.

With Grok 4.5, Claude Opus, and GPT-5.6 now competing across performance, developer adoption, and pricing, the next phase of the AI race is shifting from benchmark leadership to proving which models deliver the greatest value in real production environments.