SpaceXAI debuts Grok 4.5 model for coding, agentic AI tasks
What's the story
Elon Musk's AI venture, SpaceXAI (formerly known as xAI), has unveiled its latest artificial intelligence model, Grok 4.5. The new offering is a direct competitor to OpenAI's GPT-5.6 and Anthropic's Claude Opus 4.7 models, promising better performance and faster response times at a lower cost. The launch comes amid an intense race in the AI industry with new models being introduced almost every other week.
Model capabilities
Grok 4.5 more efficient than other leading AI systems
SpaceXAI describes Grok 4.5 as its most advanced AI model to date, specifically designed for software development and AI agents. These agents can perform complex tasks with minimal human intervention. The company claims that the new model is more efficient than other leading AI systems, using tokens twice as efficiently as other leading AI systems, which could cut down operational costs in real-world applications.
Training details
Model trained on tens of thousands of NVIDIA GPUs
The latest model has been trained on tens of thousands of NVIDIA GB300 GPUs, with a strong emphasis on improving the quality of training data by eliminating duplicates and filtering out low-quality content. It is available through SpaceXAI's coding platform, Grok Build, and can also be accessed via an API key through the SpaceXAI Console. The model has been developed in collaboration with Cursor, an AI coding start-up that SpaceX acquired in a massive $60 billion all-stock deal in June.
Pricing strategy
Grok 4.5 more affordable option compared to Anthropic's Opus 4.7
SpaceXAI has priced Grok 4.5 at $2 per million input tokens and $6 per million output tokens, making it a more affordable option compared to Anthropic's Opus 4.7, which costs $5 per million input tokens and $25 per million output tokens. The launch of Grok 4.5 comes as OpenAI gears up to publicly launch its GPT-5.6 family of models after a brief delay due to US government national security reviews.