GPT-5.6 Detects 23 Out of 26 CVEs

Alibaba's Qwen 3.8-Max showed cost competitiveness in vulnerability detection tests, but the results of this model were not included in Aikido's report. Qwen 3.8-Max achieved similar performance to competing models with a cost of $821 over three test runs. Aikido tested 13 AI models against 26 CVEs in a benchmark released on July 16. GPT-5.6 detected 23 out of 26 CVEs, achieving a recall rate of 88.5%, marking the highest performance. Claude's family detected between 15 and 18 CVEs, while Grok-4.5 detected 20. The newly added Kimi K3 also found 23 CVEs, matching the performance of GPT-5.6. This test was conducted within a code analysis harness designed for AI to explore actual code repositories and was evaluated using the 'pass@3' method. Aikido reported a total parameter count of 2.4 trillion for Qwen 3.8-Max, but the number of active parameters was not disclosed. The Aikido benchmark targeted CVEs in general software, making it difficult to interpret its performance in blockchain security audits.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.