← Back to community benchmarks

MiniMax-M2.7-oQ3

M3 Ultra (80c) · 512 GB · 3bit · 2026-04-25
Performance
32k
tokens
447.4
PP tok/s
16.4
TG tok/s
73238
TTFT (ms)
103.0
Peak mem (GB)
Hardware
Chip M3 Ultra (80c)
Memory 512 GB
GPU Cores 80
Software
oMLX v0.3.6
macOS macOS 26.2
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 642.0 44.4 95.3 GB view
4k 773.9 39.8 96.3 GB view
8k 719.2 33.5 97.2 GB view
16k 607.1 25.3 99.2 GB view
32k 447.4 16.4 103.0 GB current
Batching Results
Batch Size TG tok/s Speedup
44.4 1.00×
66.2 1.49×
95.7 2.16×
127.7 2.88×