Kimi K2.5 Achieves 4 Tokens per Second on RTX 3060 with Intel Optane
The Kimi K2.5 system operates on an RTX 3060 with 768GB Intel Optane memory, generating 4 tokens per second, despite recent server connectivity challenges.
More from this archive
The Kimi K2.5 system operates on an RTX 3060 with 768GB Intel Optane memory, generating 4 tokens per second, despite recent server connectivity challenges.