Bridgewater Associates, in collaboration with Thinking Machines Lab, has developed a Qwen3-235B model that reportedly achieves an accuracy of 84.7% in financial tasks, surpassing established models like Gemini, Claude, and GPT at a fraction of the cost. However, it is crucial to note that these results have not been independently verified, raising questions about their reliability. The absence of publicly available correct answers further complicates the assessment of these findings. This development highlights the ongoing competition in the AI landscape, particularly in the financial sector, where accuracy and cost-effectiveness are paramount. As the market evolves, understanding the implications of these results will be essential for stakeholders, particularly in light of potential regulatory changes and the need for transparency in AI performance metrics.