SpaceXAI Launches Grok 4.5, Celebrated by Elon as an ‘Opus-Class Model’

SpaceXAI has introduced its latest model, Grok 4.5 — the first launch since its public offering a few weeks ago.

In a blog post published on Wednesday, SpaceXAI characterized its new product as a powerful tool designed to tackle numerous tasks aimed at automation in the AI industry, such as coding, app building, administrative responsibilities, research, writing, and a variety of routine knowledge tasks.

According to SpaceXAI, Grok offers “double the token efficiency” compared to leading models, which may lower user costs. If this efficiency successfully translates into practical applications, it could provide a notable competitive edge for SpaceXAI, especially as token prices continue to rise for AI users.

On Wednesday, the firm published benchmark data indicating that Grok competes well with other top models from competitors, though it slightly misses the mark for best-in-class:

In a post on his social media platform X (a subsidiary of SpaceXAI), founder Elon Musk compared the new model to Opus, Anthropic’s LLM crafted for complex tasks.

“Following overwhelmingly positive feedback from our beta testing participants, @SpaceXAI will publicly launch Grok 4.5 tomorrow. It’s an Opus-class model that is faster, more token-efficient, and lower in cost,” Musk stated in a post on X.

Musk further noted: “Our internal assessments reveal that Grok 4.5 is approximately comparable to Opus 4.7 but significantly faster. This combination of capability, speed, and cost-effectiveness provides it with a competitive advantage.”

As indicated by SpaceXAI, Grok 4.5 is priced at $2 per million input tokens and $6 per million output tokens. This pricing is highly competitive, assuming Grok’s performance meets SpaceXAI’s assertions.

In contrast, Opus 4.7 charges $5 per million input tokens and $25 per million output tokens. OpenAI offers various pricing structures for its different model iterations, with Sol, the priciest option, costing $5 for input tokens and $30 for output, while its most affordable model, Luna, comes in at $1 for input and $6 for output.

This week holds considerable significance for AI model launches. On Thursday, OpenAI plans to unveil GPT 5.6, its latest and most potent model, which had previously faced restrictions from the Trump administration due to security concerns. OpenAI has labeled it as its “most advanced model to date.”