Mercury 2.5 LLM hits 770 tokens per second
Mercury 2.5, a model priced moderately among its peers, has demonstrated impressive performance with a token output rate of 770 tokens per second. This model operates within a 260k token context window and has earned a score of 12 on the Artificial Analysis Intelligence Index, placing it slightly below the median of 13. Despite its average intelligence, Mercury 2.5 stands out for its speed and conciseness.
It is capable of generating 35 million tokens, which is comparatively concise to the median of 85 million. The model's pricing, both for input and output tokens, aligns with the median values, making it cost-effective for users. With an Elo score of 500 and a weighted average cost per task, Mercury 2.5 proves to be a valuable tool for various applications, particularly in Agentic real-world work tasks.
Written by urgent.news from Hacker News's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
