Elon Musk’s xAI has officially released Grok 4.7, marketing it as the company’s most capable model to date for coding and knowledge work. According to the company, Grok 4.7 is built on a larger base model, utilizes longer reinforcement learning, and features enhanced self-verification capabilities. xAI has aggressively priced the model at $2 per million input tokens and $6 per million output tokens, positioning it closer to low-cost Chinese models than Western frontier competitors.
Despite the competitive pricing, benchmark performance indicates a noticeable capability gap compared to rival models. On the independent Artificial Analysis Intelligence Index (v4.3.2), Grok 4.7 scored 46, placing it mid-pack behind Claude Fable 5.1 and GPT-6, which led with scores of 53. The disparity is particularly pronounced in agentic coding tasks on Terminal-Bench 4.0, where Grok 4.7 achieved a 26 percent success rate, trailing GPT-6 Astra at 60 percent, Claude Fable 5.1 at 55 percent, and DeepSeek V4.1 Flash at 27 percent.
The new model is currently accessible to developers via the Grok API, Cursor, and Grok Build. The pricing strategy suggests xAI is attempting to drive developer adoption through cost efficiency while working to bridge the performance gap on complex autonomous tasks.
Why it matters
Offers a cost-effective alternative for routine coding and knowledge tasks, though performance lags on complex agentic workflows.
Highlights growing price competition as xAI undercuts Western rivals to match Chinese model economics.
Signals that self-verification and longer reinforcement learning training have not yet closed the benchmark gap with top-tier models.
Source: the-decoder.com
