According to the company, the model’s “logical reasoning” capability has outperformed Claude Opus 4.1 in tests and is on par with Gemini 2.5 Pro.
The Grok 4 Fast model supports a context size of 2 million tokens and is available for free on Grok’s website as well as iOS and Android apps. According to tests conducted by Artificial Analysis, Grok 4 Fast even surpassed OpenAI’s GPT-5 and o3 models in data retrieval mode.
Researchers have observed that xAI’s new model uses tokens more efficiently than its competitors. In tests, Grok 4 Fast required 61 million tokens to complete all tasks, while Gemini 2.5 Pro used 93 million tokens. Thanks to this efficiency, using Grok 4 Fast via API is 23 times cheaper than GPT-5 and 25 times cheaper than Gemini 2.5 Pro.
API pricing is as follows: $0.2 per 1 million input tokens and $0.5 per 1 million output tokens.
Grok 4 Fast is available for a limited-time free trial on the OpenRouter and Vercel AI platforms. Users are already testing the model’s coding capabilities and obtaining comparative results. For example, one user submitted a query to Grok 4 Fast and GPT-5 to write an interactive website code on the “Harry Potter” theme, showing that Grok produced more complete and functional code.
