Created bymario-alka.dePowered bynoob2claw.detricoma.de
Manufacturer

MiniMax

MiniMax-M-Modellfamilie.

6Models
158Benchmarks
144Performance
14Harness
Best token generation
551,0 tok/s
MiniMax-M2.5
Best prefill
7.167,2 tok/s
MiniMax-M2.7-AWQ-4bit
∅ Token generation
77,4 tok/s
Average of 144 runs
∅ Prefill
1.482,1 tok/s
Average of 144 runs
Best TTFT
105 ms
MiniMax-M2.7-AWQ-4bit
Best harness rate
81%
minimax-m2.5-awq
🏆
MiniMax-M2.5 · fastest model at 551,0 tok/s token generation
Best prefill of this manufacturer: 7.167,2 tok/s (MiniMax-M2.7-AWQ-4bit)

Top models by token generation

Best measured generation throughput per model (tok/s), published performance benchmarks.

MiniMax-M2.5
551,0 tok/s
MiniMax-M2.7
548,6 tok/s
minimax-m2
547,9 tok/s
MiniMax-M2.7-AWQ-4bit
108,6 tok/s
minimax-m2.5-awq
105,6 tok/s