SpaceXAI released Grok 4.7, an AI model for cybersecurity that achieved a composite score of 56 on the Artificial Analysis Cyber Index, tying with MiMo-V2.6-Pro for first place in the ranking. This benchmark measures AI agents’ ability to defend software by identifying vulnerabilities, reproducing them, and shipping working patches within real codebases. Grok 4.7 posted a 74% success rate on the CyberGym-E2E-AA test and a 68% pass@1 rate on CWE-Bench-AA. The model carries a high cost of $11.67 per task compared to cheaper competing models. Grok 4.7 comes from SpaceXAI, the entity formed after the merger with xAI, and Elon Musk highlighted the ranking on social media.
Source: Read the original article
Disclaimer: this content is for information purposes only and is not financial advice. Cryptocurrencies are highly volatile: you may lose all of your capital. Always do your own research. Legal notice
