Highly variable token/sec for minimax m3

beasleypero June 20, 2026, 5:15am 1

Sometimes I get ~70 tok/sec other times I get ~3 tok/sec, is there any way to make this more stable?

Topic		Replies	Views
MiniMax-M3 (428B MoE + vision) at ~14–15 tok/s on 2× DGX Spark — EAGLE3 speculative decoding is the unlock DGX Spark / GB10 Projects llama , deepseek	2	786	July 4, 2026
Minimax3 on 2 nodes decode ~10.7 tok/s, 4bits DGX Spark / GB10 llama	26	1617	June 21, 2026
MiniMax M3 : NVFP4 for Quad DGX Spark DGX Spark / GB10 agentic-ai , deepseek	116	8514	June 25, 2026
MiniMax M3 NVFP4 and NVFP4 REAP 50 for 4x & 2x DGX Sparks DGX Spark / GB10 Projects	53	4297	July 2, 2026
Minimax-m3 internal server error Models nim	8	725	June 24, 2026
MiniMax-M3-AWQ on 4× GB10, fp8 KV, 262k context, adaptive reasoning, ~30 tok/s DGX Spark / GB10 llama	17	1264	July 21, 2026
Getting "Too Many Requests" always in MiniMax M3 Models	3	273	August 10, 2026
Many models are very very slow to start answer or thinking NVIDIA Nemotron agentic-ai	0	612	April 26, 2026
Minimax-m2.7 Models nim	2	174	July 28, 2026
MiniMax M2.5 released (not available on HuggingFace as of now) -- is DGX Spark ready? DGX Spark / GB10	92	6921	April 12, 2026