Grok 4.5 Challenges OpenAI And Anthropic With Cheaper Agentic AI

SpaceXAI launched Grok 4.5 as a lower-cost AI model for coding and agentic work after its $60B purchase of Cursor.

Key Points:

Grok 4.5

SpaceXAI released Grok 4.5 on Jul. 9, presenting the model as an early result of Elon Musk’s push to build a vertically integrated AI operation. The launch follows the company’s $60B acquisition of Cursor, an AI coding tool that gives SpaceXAI both developer workflow data and a potential distribution channel.

The company said Grok 4.5 is built for coding, autonomous agent tasks and enterprise workloads that require repeated tool calls, long code reviews and iterative development. It priced the model at $2 per million input tokens and $6 per million output tokens, below premium models from Anthropic and OpenAI.

Musk told staff the model is roughly comparable to Claude Opus 4.7, but faster. Artificial Analysis ranked it fourth on an agentic knowledge work index, while also estimating that its cost per completed task was nearly 90% lower than higher-ranked alternatives.

That is the core pitch.

Also Read: Bitcoin’s Recovery Has A Demand Problem Bulls Cannot Ignore

Coding AI

The launch targets companies that run AI agents across engineering teams, where token use can rise quickly during code analysis and production debugging. In those settings, a small performance gap may matter less if the model completes large volumes of work at a much lower cost.

Early feedback from Tesla and SpaceX engineers suggests Grok 4.5 has been useful in production settings, especially with large codebases spread across several repositories. Cursor’s interaction data reportedly helped train the model on real developer behavior, which can differ from benchmark tasks.

The release also comes after earlier Grok models drew criticism for antisemitic outputs and unsafe image generations that attracted regulatory scrutiny. Several co-founders left earlier this year, and Musk later acknowledged deeper issues before starting a rebuild.

SpaceXAI’s strategic advantage is the closed loop around compute, models, training data and internal demand. The Colossus supercomputer cluster supports training, Cursor supplies specialized coding data, and Tesla and SpaceX give the company large real-world engineering environments.

The risk is that coding quality depends on reliability that benchmarks often miss. A cheaper model can lose its advantage if developers need more retries, manual checks or corrections to reach the same result.

The broader market has moved toward practical AI economics as much as raw intelligence scores. Grok 4.5 reflects that shift, with SpaceXAI betting that cost and speed can become decisive for enterprise coding agents.

Read Next: Polymarket Traders Back Avengers Doomsday To Crush Dune 3 Opening Weekend