According to CoinMeta, Qianwen’s three-round vulnerability detection recall rate reached 81.3%, tying for first place with Opus. Its total cost was approximately $821, which is only half of that of Opus and SOL. Security company Aikido tested 7 models using code auditing, including Qwen3.8-max, Qwen3.8-max, Kimi K3 Max, DeepSeek V4 Flash, as well as GPT-5.6, SOL, Luna, and Terra. The tests covered 32 recently disclosed vulnerabilities, with each model being run three times. Qwen3.8-max identified a total of 26 vulnerabilities across the three rounds, with a comprehensive score of 83.2%, which is slightly lower than that of Opus, Kimi K3 Max, GPT-5.6, and SOL. However, Qianwen’s performance is not consistent; among the 26 vulnerabilities identified, only 10 were detected in all three rounds, whereas both Opus and SOL detected 19 vulnerabilities in each round.