排序
异议!顶流AI决战「逆转裁判」:o1险胜Gemini 2.5登顶、Llama 4零分垫底
UCSD研究团队以经典游戏《逆转裁判》为舞台测试AI推理能力,结果显示o1和Gemini 2.5 Pro表现最佳;通过项目开源,更多经典游戏可用来测试AI模型性能。
Ilya新公司估值320亿美元,没有任何产品
SSI (Safe Superintelligence) has raised $2 billion and is valued at $32 billion, with plans to focus on safety and innovation rather than immediate products or cycles.