Data compiled by Woofun AI shows that MindScape AI's MoE flagship model GLM-5.2 has secured the top position in the Artificial Analysis Big Model Intelligence Index v4.1, achieving a score of 51. This performance surpasses competitors including MiniMax-M3 and DeepSeek V4 Pro, while its GDPval-AA v2 score of 1524 ties with the closed-source GPT-5.5, significantly outperforming the human baseline of 1000.

The model architecture comprises 744B total parameters with 40B active parameters and an expanded context window of 1M tokens under the MIT license. While the average single-task cost has risen to approximately $0.46 due to higher token output, it remains substantially lower than equivalent closed-source alternatives. Services are now available via the MindScape official API and platforms such as SiliconFlow, DeepInfra, and Nebius AI.