SpaceXAI releases Grok 4.5, which Elon describes as an ‘Opus-class model’
Key Takeaways
- SpaceXAI, formerly known as xAI, released Grok 4.5 on Wednesday, describing it as an “Opus-class model,” according to TechCrunch.
- SpaceXAI was acquired by SpaceX in February, according to TechCrunch, meaning Wednesday’s release is not tied to any recent public listing of the company.
- Grok 4.5 is priced at $2 per million input tokens and $6 per million output tokens, according to TechCrunch, compared to Anthropic’s Opus 4.7 at $5 per million input tokens and $25 per million output tokens, though sources note a new tokenizer in Opus 4.7 may complicate direct price comparisons.
- OpenAI announced the global, public availability of its GPT-5.6 model series starting Thursday, July 9th, according to TechCrunch.
What SpaceXAI Announced
SpaceXAI, the company formerly known as xAI that was acquired by SpaceX in February, used a blog post on Wednesday to introduce Grok 4.5, positioning it as a general-purpose tool capable of handling coding, app development, administrative tasks, research, and writing, according to TechCrunch. The company framed the release as its answer to the kind of routine knowledge work that rival AI labs have also targeted with their own flagship systems.
A central part of SpaceXAI’s pitch is efficiency. The company claims Grok 4.5 offers twice the token efficiency of other leading models, according to TechCrunch. Token efficiency refers to how much computational output a model can generate relative to the resources it consumes, and it has become an increasingly important metric as businesses weigh the ongoing costs of running AI systems at scale rather than just their upfront capabilities.
SpaceXAI also published benchmark results on Wednesday that, according to TechCrunch, showed Grok 4.5 performing competitively against other top-tier models, though not quite matching the very best scores across the board. This positions the release as a strong contender rather than an outright category leader on raw performance alone.
Musk, in a post on X, the social media platform that TechCrunch notes is a subsidiary of SpaceXAI, wrote that based on strong positive feedback from customers in the company’s beta test program, SpaceXAI would make Grok 4.5 available to the public the day after the announcement. He added that the model is “an Opus-class model, but faster, more token-efficient and lower cost.” In a follow-up post, Musk said the company’s internal assessment placed Grok 4.5 as roughly comparable to Opus 4.7, but notably faster, and that the combination of capability, speed and lower cost is what he believes makes it competitive.
Why the Pricing Details Matter
The numbers SpaceXAI put forward are central to its argument. Grok 4.5 is priced at $2 per million input tokens and $6 per million output tokens, according to TechCrunch. By comparison, Anthropic’s Opus 4.7 costs $5 per million input tokens and $25 per million output tokens, according to TechCrunch, though industry sources indicate that a new tokenizer introduced with Opus 4.7 may affect how directly its pricing can be compared to other models, since token counts themselves can vary between systems.
OpenAI’s pricing varies by tier: its most expensive model, Sol, costs $5 per million input tokens and $30 per million output tokens, while its cheapest option, Luna, costs $1 per million input tokens and $6 per million output tokens, according to TechCrunch.
For businesses and developers who build products on top of large language models, these per-token costs directly affect operating margins, particularly for applications that generate large volumes of output, such as coding assistants or research tools. A model that claims comparable capability at a lower output cost could reshape which providers companies choose to build around, especially as usage scales.
The release also arrives in a crowded week for major model launches. OpenAI announced the global, public availability of its GPT-5.6 model series starting Thursday, July 9th, according to TechCrunch, calling it the company’s “strongest model yet.” According to TechCrunch, the release of that model had previously been limited due to concerns raised by the Trump administration about its security implications, though it is now being made broadly available.
Hype Check
Claim: SpaceXAI says Grok 4.5 matches the capability of Anthropic’s Opus-class models while costing significantly less per token, and that it is roughly comparable to Opus 4.7 but faster, according to statements from Elon Musk cited by TechCrunch.
Reality: SpaceXAI’s own benchmark data, as reported by TechCrunch, shows Grok 4.5 performing competitively but just short of the best-in-class scores among top models. The pricing comparison with Opus 4.7 is also complicated by reports of a new tokenizer that may affect how token costs are measured across systems, meaning the headline price gap may not translate directly into real-world savings.
Verdict: Grok 4.5 appears to be a credible, cost-conscious entrant into the crowded field of high-end AI models, but claims of matching Opus-class performance at a fraction of the cost should be weighed against the caveats around benchmark scope and tokenizer differences. As with any fast-moving corner of the AI industry, independent verification over time will matter more than launch-day comparisons. This is not financial advice.
Source
Researched with AI assistance, fact-checked and edited by a human. Not financial advice.