The AI market has become lively once again as SpaceXAI introduces its latest strategic product, Grok 4.5. This is also the first model released by the company since its high-profile IPO a few weeks ago.

According to SpaceXAI, Grok 4.5 is positioned as a versatile “workhorse” designed to smoothly handle familiar tasks that the AI industry is trying to automate: from coding and app development to office work, research, writing and other everyday intellectual tasks.

Twice the Token Efficiency and Surprisingly Low Pricing

The most notable point of Grok 4.5 is its cost structure. SpaceXAI claims the new model offers “twice the token efficiency” of today’s leading competitors. If this is proven in real-world use, it could become a major advantage at a time when token costs remain a serious concern for both companies and individual users.

 

In terms of specific pricing, SpaceXAI announced that Grok 4.5 costs $2 per million input tokens and $6 per million output tokens.

To better understand how aggressive this pricing is, it is worth comparing it with competitors:

  • Anthropic’s Opus 4.7 costs up to $5 per million input tokens and $25 per million output tokens.
  • OpenAI divides its models into several tiers: the premium Sol line costs $5 for input and $30 for output per million tokens, while the lower-cost Luna line is more affordable at $1 for input and $6 for output per million tokens.

Clearly, if Grok 4.5’s real-world capabilities match the claims, this pricing could create enormous competitive pressure on rival companies.

Elon Musk: “Grok 4.5 Is at Opus 4.7 Level, but Faster”

On X, which is a subsidiary of SpaceXAI, Elon Musk strongly praised the company’s new flagship model. He compared Grok 4.5 with Anthropic’s Opus line, which is known for handling heavy and complex tasks.

“Based on very positive feedback from customers in the beta testing program, @SpaceXAI will officially open Grok 4.5 to the public tomorrow. This is an Opus-class model, but faster, more token-efficient and less expensive,” Musk said enthusiastically.

He then added: “Our internal evaluations show that Grok 4.5 has capabilities equivalent to Opus 4.7, but with much faster processing speed. The combination of strong performance, superior speed and lower cost is our competitive weapon.”

The published benchmark figures also show that Grok 4.5 is fully capable of competing directly with today’s top models, although it still falls slightly short of reaching the best-in-class position.

A Tense Week for AI Giants

The launch of Grok 4.5 at this moment looks like a calculated pre-emptive strike. Just one day later, OpenAI is also scheduled to release GPT-5.6, the company’s most powerful model to date, which had previously been delayed by the Trump administration due to security concerns.

The AI race is clearly entering a more intense phase than ever, as major players rapidly release new models, competing not only on intelligence but also by aggressively lowering prices to capture market share.

Grok 4.5 shows that the AI race is shifting from pure model intelligence to a broader battle over speed, pricing and token efficiency. If SpaceXAI can deliver Opus-level performance at a lower cost, it may force major AI providers to rethink their pricing strategies.