Elon Musk’s AI company, XAI, starts an API for Grok 3

Billionaire Elon Musk has just been reversed by Openai, but this does not stop the company AI, XAI, by Making the Grok 3 model available via API.

It’s been several months since Xai revealed Grok 3, the company’s response to models such as Openai’s GPT-4O and Google’s Gemini. Grok 3 can analyze the images and respond to questions and authorize a series of features on the social network of Musk, X, which was not so accidentally acquired by XAI in March.

XAI offers two flavors of the flagship model through API: Grok 3 and Grok 3 Mini with “reasoning” capabilities.

Grok 3 is priced at $ 3 per million tokens (~ 750,000 words) fueled in the model and $ 15 per million brands produced by the model. Meanwhile, the Grok 3 mini will cost $ 0.30 per million input brands and $ 0.50 per million tokens.

The fastest versions of the Grok 3 and the Grok 3 mini are available with a premium: $ 5 per million input brands and $ 25 per million tokens for Grok 3? and $ 0.60 per million input brands and $ 4 per million production brands for Grok 3 mini.

Grok 3 is not cheap in relation to competition. The XAI fits Anthropic’s Claude 3.7 Sonnet pricing, which also offers reasoning capabilities, but is more expensive than the recently released Google Gemini Pro, which generally achieves higher ratings than the popular AI index. (Specifically Xai was accused of misleading Grok reference reports 3.)

As Several users in x pointed outThe Grok 3 through XAI’s API also has a smaller environment window than the model supposed to be capable of supporting. (The “Environment window” refers to how many brands the model can be processed with a Go.) API MAXES in 131,072 chips or about 97,500 words – except for 1 million XAI brands claimed that Grok 3 is supported in late February.

When Musk announced Grok about two years ago, he threw AI as intense, non-filtered and anti-“awakened”-generally, willing to answer controversial questions, other AI systems will not. Gave some of these promise. He said to be vulgar, for example, the Grok and Grok 2 would be fortunately forced, hitting a colorful tongue that you probably won’t hear from Chatgpt.

But the Grok models before Grok 3 winged on political issues and will not cross certain limits. Indeed, a study He found that Grok leaned into political left on issues such as transsexual rights, diversity programs and inequality.

Musk blamed this behavior in Grok’s training data – public websites – and committed To “shift Grok closer to politically neutral.” With no high profile mistakes, as he briefly censors the unexpected reports of President Donald Trump and Musk, it is not yet clear whether Xai has achieved this goal at the model level and what are the long -term consequences.