MERCURY 2.5 LLM REACHES 770 TOKENS PER SECOND

Mercury 2.5, a new large language model, has achieved a throughput of 770 tokens per second. The benchmark represents a significant performance milestone in LLM inference speed.