AI model · by xAI
Grok 4.7
Grok 4.7 is xAI's top-tier model, aimed at coding and broad work, with support for agentic tool use and adjustable reasoning.
- 500K tokens context
- Closed weights
- Checked against xAI's page 19 Oct 2026
What Grok 4.7 is for
xAI says Grok 4.7 is its leading model for programming and everything else, offering agentic tool use, minimal hallucinations, and adjustable reasoning. xAI also lists a 500k-token context and a knowledge cut-off of May 2026.
What xAI says it does well
- Agentic tool calling
- Minimal hallucinations
- Configurable reasoning
- Context window of 500k tokens
Limits xAI mentions
- Has no knowledge of current events or data beyond its training data unless server-side search tools are enabled
- Knowledge cut-off date is May 2026
- Planned: logprobs and top_logprobs are not supported; fields are silently ignored if set