MiniMax M3 Closes Gap to GPT-4o Performance on AA-Briefcase

The MiniMax M3 AI model has made significant progress in its testing on the ArtificialAnlys AA-Briefcase, surpassing the performance of 4.6 Max and closing in on GPT-4o's capabilities. After two weeks of DeFi front office testing, MiniMax M3 now ranks #5 on the AA-Briefcase.

The model's 1M context window and multimodal edge are seen as key factors contributing to its success, with some traders considering it a viable alternative alongside GLM5.2. However, even after rework, only 70 percent of outputs from MiniMax M3 reach top-tier usability.

ManusAI 1.6 Max remains the stronger commercial pick due to near-100 percent usable results and 20-30x immediate value over standard 4.8 API pricing. This highlights the ongoing challenge in finding reliable frontier tools for AI adoption in DeFi front office trading desks.