xAI launches Grok 4.7 at bargain prices, but benchmarks reveal a wide gap to Claude and GPT-6
decoder ·
xAI has released Grok 4.7, positioning it as its most advanced model for coding and knowledge work, built on a larger base with extended reinforcement learning. The company offers the model at highly competitive rates of $2 per million input tokens and $6 per million output tokens, aiming to rival Chinese pricing structures. However, independent benchmarks from Artificial Analysis reveal that Grok 4.7 scores only 46 out of 100, placing it significantly behind leaders like Claude Fable 5.1 and GPT-6, which both scored 53. The performance gap becomes even more pronounced in agentic coding tasks, where Grok 4.7 achieved just 26 percent on Terminal-Bench 4.0. In contrast, GPT-6 Astra reached 60 percent and Claude Fable 5.1 reached 55 percent on the same test. Notably, the cheaper DeepSeek V4.1 Flash also slightly outperformed Grok 4.7 with a score of 27 percent. Despite these shortcomings relative to Western frontier models, Grok 4.7 is currently accessible via the Grok API, Cursor, and Grok Build platforms.
decoder