llamaperf
Sep 23, 2026
Throughput
68.3 t/s gen

Summary

User reports MiMo V2.6 Pro RL at 68.3 t/s on eight DGX Spark units, up from 17.8 t/s without speculation. Setup is vLLM with DFlash speculative decoding on official weights; the figure includes prefill and request overhead. Outputs matched byte for byte across 12 cases, including two baseline mistakes. Four concurrent requests and a 257K-token input were also tested.