SpaceXAI Grok 4.5 release: price, benchmarks, Musk view

SpaceXAI has released Grok 4.5, the company’s first model launch since it went public several weeks ago, announced in a blog post published Wednesday. The company positions Grok 4.5 as a broad workhorse for coding, app-building, office and clerical work, research, writing, and routine knowledge work.

What does Grok 4.5 claim to do?

Grok 4.5 is presented as an all-purpose model for coding, app-building, office and clerical work, research, writing, and other routine knowledge tasks, and SpaceXAI says it is more token-efficient than competitors. The company described the release as a workhorse that can tackle the typical tasks the AI industry automates, and it claimed “twice greater token efficiency” compared with other leading models.

The firm also published benchmark metrics that the company says show Grok’s competitiveness with other top models, though those metrics appeared to place Grok just short of best-in-class. Founder Elon Musk compared the model to Anthropic’s Opus line on his social platform, calling it an "Opus-class model" and saying their internal assessment was that "Grok 4.5 is roughly comparable to Opus 4.7, but much faster." Musk added that faster speed plus lower cost and capability is the combination that makes it competitive.

How does Grok 4.5’s pricing and efficiency compare to rivals?

SpaceXAI set explicit token prices for Grok 4.5 at $2 per million input tokens and $6 per million output tokens. By comparison, the company noted Opus 4.7 costs $5 per million input tokens and $25 per million output tokens. OpenAI’s pricing tiers were also cited: Sol costs $5 per million input and $30 per million output tokens, while Luna costs $1 per million input and $6 per million output tokens.

Those published prices illustrate where SpaceXAI is positioning Grok 4.5: lower per-token costs than Opus 4.7 and OpenAI’s most expensive tier, and competitive with lower-cost tiers on output pricing. SpaceXAI pairs that pricing with its claim of "twice greater token efficiency," which, if borne out in practice, would reduce real usage costs for consumers where token consumption dominates billings.

The company also released benchmark materials alongside the launch. SpaceXAI presented metrics intended to show Grok’s performance versus other leading models, and the coverage noted the results suggested strong competitiveness while falling short of the top performers on some measures.

Why it matters

Lower token costs combined with improved token efficiency directly address a growing concern for AI consumers: the expense of tokens. If Grok 4.5 truly delivers lower per-token billing and uses tokens more efficiently, organizations that run large volumes of requests could see materially lower bills. Speed, which Musk singled out, matters for interactive applications and high-throughput workloads, and SpaceXAI is pitching that combination of capability, speed, and lower cost as its competitive angle.

What to watch

Watch for independent, real-world comparisons of Grok 4.5’s token efficiency and latency, and whether the company’s benchmark claims hold up outside its materials. Also monitor competing launches: the article notes OpenAI planned to release GPT 5.6 on Thursday, which will be another point of comparison for capability and cost.

SpaceXAI’s public release of Grok 4.5 comes as the company emphasizes pricing and efficiency as selling points; the next validation will be user feedback and third-party benchmarks that measure cost per task, not just token rates.