1M Tokens/s: Scaling Qwen 3.5 27B on 96 B200 GPUs with vLLM

3 pointsposted 6 hours ago
by m4r1k
(medium.com)

No comments yet