lon Musk's xAI artificial intelligence company has just made public that its Grok 3 model will be available for users via an API. Elon Musk launched the Grok 3
model as an industry response to maintain its authority in the AI field among impressive competitors such as GPT-4o and Gemini models.
Grok 3, developed by the xAI artificial intelligence company, is able to respond to users’ questions and analyze different images. Also, this model is used to power the social media network, X (formerly Twitter), which is also owned by Elon Musk. Coincidence or not, this popular social media platform was recently purchased by xAI in March, which raised the platform’s valuation to $33 billion.
It should be mentioned that the xAI company offers two options, Grok 3 and Grok 3 mini, which have reasoning capabilities. Grok 3 has a price of $3 for a million tokens, which is equivalent to 750,000 words that are entered into the AI, and a price of $15 for a million tokens generated by the Grok 3 model.
On the other hand, the Grok 3 mini has a significantly lower cost, of $0.30 per million entered tokens and $0.50 per million generated tokens.
Users should know that these two models are also available in premium versions. Grok 3 premium version costs $5 per million entered tokens, with a generated million tokens of $25. With a much lower cost, Grok 3 mini premium versions require $0.60 for a million input tokens, while for the million output tokens, Grok 3 mini requires $4.
Subscribe to our newsletter
Compared with the other industry models, the xAI Grok 3 model is not really cheap, but it seems that the AI company wants to match Anthropic’s pricing range available for Claude 3.7 Sonnet. It’s true that Claude 3.7 Sonnet offers similar reasoning capabilities, but has a higher price than Gemini 2.5, which is able to record much higher performance scores than Grok 3 in most artificial intelligence benchmarks.
As several X users have reported, compared with the industry artificial intelligence models, Grok 3 has a significantly smaller context window. Context window represents the number of tokens that a model can process in just one input. So, it seems that the new Grok 3 API is limited to 131,072 tokens, equivalent to 97,500 words, which is a significantly lower amount than the 1 million tokens claimed by the xAI company a couple of months ago.
Is the new Grok 3 API able to satisfy the industry’s needs? Stay tuned to find out the Grok 3 performances compared with other artificial intelligence models.
.webp)
.webp)
