On Wednesday local time, Elon Musk's SpaceX AI officially launched its latest large model, Grok 4.5, marking the company's first major product update since its listing weeks ago. Musk personally promoted it on the social platform X, comparing Grok 4.5 to Anthropic's high-end model Opus, designed for complex tasks, and called it "an Opus-level model, but faster, more token-efficient, and lower in cost."

image.png

According to a blog post released by SpaceX AI, Grok 4.5 is positioned as a "mainstay" for general scenarios, capable of handling various tasks currently at the forefront of AI automation, such as code writing and application development, office and document processing, research and writing. Benchmark test data shows that its overall capabilities have entered the top tier, slightly below the current "ceiling-level" models, but still within a highly competitive range.

Token efficiency doubled, pricing strategy highly aggressive

A key selling point of Grok 4.5 is "double the token efficiency," meaning fewer tokens are generated under the same workload, significantly reducing inference costs. In the current context where AI model call prices have become a core consideration for enterprise users and developers, if this efficiency advantage is verified in practical scenarios, it will become a significant bargaining chip for SpaceX AI competing with top models.

At the pricing level, SpaceX AI has set an aggressive price: $2 per million input tokens and $6 per million output tokens. For comparison, Anthropic's Opus 4.7 charges about $5 per million input tokens and $25 per million output tokens; OpenAI's most expensive Sol version charges $5 for input and $30 for output. Even compared to OpenAI's entry-level Luna model, which charges $1 for input and $6 for output, Grok 4.5 matches only on the output side, but its comprehensive capabilities are clearly far ahead.

This week, AI products are being released intensively, and the model competition is full of tension

Musk revealed that based on strong positive feedback from trial plan customers, Grok 4.5 will be fully opened to the public the next day. He added that according to internal assessments, Grok 4.5's comprehensive capabilities are "roughly comparable to Anthropic's Opus 4.7, but with faster speed," and pointed out that it is the combination of "capability, speed, and cost" that forms the core competitiveness of this model.