Mercury 2.5 LLM hits 770 tokens per second

AI Digest

Mercury 2.5 LLM 以780 tokens/s速度和合理价格脱颖而出，但智能指数低于平均水平，适合高效文本处理场景。

Mercury 2.5 achieves 780 tokens/s speed and moderate pricing, but its intelligence index lags, making it suitable for high-efficiency text tasks.

Key points

• 智能指数12分，低于同类模型中位数13分 Intelligence Index score of 12, below median of 13 among comparable models
• 速度达780.8 tokens/s，排名前112位 Speed of 780.8 tokens/s ranks top 112
• 输入0.25美元/1M token，输出0.75美元/1M token Moderate pricing at $0.25 input, $0.75 output per 1M tokens
• 支持260k tokens上下文窗口，输出简洁 260k token context window with concise output
• 同类模型中性价比高，适合高效任务 High cost-effectiveness compared to similar models

Takeaway: Mercury 2.5适合需要高效处理的场景，性价比高但智能稍逊。 / Suitable for high-efficiency text tasks with good cost-effectiveness, though intelligence lags slightly.

Why it matters 适合需要高效文本处理的场景，性价比高且速度优势明显。