SpaceXAI releases Grok 4.5 and says it will cost less to run

In short: SpaceXAI has released Grok 4.5, a new version of its Grok AI model, and says it is faster and cheaper to use than some competitors.

What happened

SpaceXAI announced Grok 4.5 in a Wednesday blog post, calling it a general purpose model for common work tasks. The company said it can help with coding, office work, research, and writing.

Elon Musk said on X that Grok 4.5 will be made available to the public the next day, after feedback from a beta test program. He compared it to Anthropic’s Opus, which is another well known AI model used for complex tasks.

SpaceXAI also shared benchmark results, which are test scores that try to measure how well an AI model performs on certain tasks (like giving students the same exam to compare grades). The company said the scores suggest Grok 4.5 is competitive with other top models, though not necessarily the best in every test.

A big part of SpaceXAI’s message is cost. It priced Grok 4.5 at $2 per million input tokens and $6 per million output tokens. Tokens are small chunks of text that AI systems count, like charging by the word, but grouped into pieces rather than exact words.

Why it matters

For people and businesses using AI tools, the bill can add up quickly, especially when the tool reads a lot of text and generates long answers. If Grok 4.5 performs similarly to other leading models but uses fewer tokens and runs faster, it could lower costs for apps and services that rely on AI.

Source: TechCrunch AI