Kimi K3 is the largest model we feature, with 2.8T total and 104B active parameters. MI355X serves ~8M tokens/s per MW at 50 tok/s per user, about a third of B300 (~24M), which uses NVIDIA's multi-node disaggregated configs. Its curve is shallow, peaking at about 13M at 12 tok/s per user, and for a single agent it reaches 152 tok/s against 235 on B300.
Kimi K3 serves 8M tokens/s per MW on MI355X, about a third of B300
AISummary
Artificial Analysis says Kimi K3, with 2.8T total and 104B active parameters, serves about 8M tokens/s per MW on AMD MI355X at 50 tok/s per user, roughly a third of the roughly 24M reached on NVIDIA B300 with multi-node disaggregated configs. The MI355X curve peaks at about 13M tokens/s per MW at 12 tok/s per user. For a single agent, MI355X reaches 152 tok/s against 235 tok/s on B300.
Post on XView on X
Source: Artificial Analysis · x.comPublished