SpaceXAI recently released a new version of its chatbot called Grok 4.5. This is a programming-focused model that Elon Musk pitched as an “Opus-class” competitor to Claude Anthropic at a fraction of the cost.
Given the increasing cost of using AI, engineering teams are now prioritizing the price per task completed over peak intelligence.
Does your new GROC model outperform the Cloud Opus 4.8?
Elon Musk described, in a post on X, Grok 4.5 is newly launched As an “Opus class model” it is faster and more cost effective. The model is designed for coding and other complex tasks.
Musk continued after the launch with a second post: writing“The Grok 4.5 appears to be #1 in at least a few benchmarks. Better than expected.”
SpaceXAI (NASDAQ: SPCX) lists Grok 4.5 at $2 per million input codes and $6 per million output codes, which is much cheaper than using Anthropic’s models, which are priced at $5 and $25 for Claude Opus 4.8, or OpenAI’s GPT-5.6 Luna, which is priced at $1 and $6.
The new model is the first to be developed jointly with Cursor, an AI coding startup, which SpaceXAI recently agreed to buy for $60 billion in an all-stock deal. The deal is expected to close in the third quarter of 2026.
Cursor said it partnered with SpaceXAI for training, feeding real-time cues on how engineers should write, review and debug. Trained across tens of thousands of Nvidia GB300 GPUs, SpaceXAI says the model runs at about 80 tokens per second with nearly twice the token efficiency of similar competitors.
Where does Grok actually rank among its competitors?
Independent comparison site Synthetic analysisGrok scored a 4.5 at 54 on its IQ, putting it in fourth place behind Anthropic’s Fable 5, OpenAI’s GPT-5.5, and Cloud Opus’s 4.8.
In coding tests, Grok 4.5’s synthetic analysis in Grok Build gave it a score of 76 in its test Coding proxy index. This is the same level as GPT-5.5 in Codex and slightly lower than Fable 5 in Claude Code.
The site priced Grok 4.5 at $2.49 per coding task compared to $11.80 for Fable 5 at Claude Code. Grok 4.5 consumed approximately 1.9 million tokens per mission, while Fable 5 used 7.2 million.
By one specific criteria, Grok had a 4.5 clean win over Claude. Snorkel AI The company ran the model against GPT 5.5 and Opus 4.8 on a sample of about 2,000 professional business tasks from its GDPval+ dataset.
Grok 4.5 passed 29% of them, ahead of GPT 5.5 at 22% and Opus 4.8 at 21%. The widest margins were in legal, educational and health care tasks. Grok 4.5 was paired with SpaceXAI’s Grok Build agent to conduct this test at the company’s request.
Cryptopolitan I mentioned Earlier in the week, Elon Musk’s statement about the model being “close to Opus and perhaps exceeding it” stemmed from the impression of the company’s evaluators rather than a published benchmark. The specific Opus the Grok was measured against was the Claude Opus 4.8.
Musk himself later admitted that the model competes with a previous generation of Claude rather than the current one.
The Grok 4.5 is available now through the SpaceXAI console, the Grok Build dealer, and in-currency on all plans, and is expected to arrive in the EU in mid-July.
Don’t just read cryptocurrency news. Understand that. Subscribe to our newsletter. It’s free.
