← Back to community benchmarks

gpt-oss-120b-MXFP4-Q8

M3 Ultra (80c) · 512 GB · 4bit · 2026-03-10
Performance
4k
tokens
693.6
PP tok/s
68.2
TG tok/s
5905
TTFT (ms)
59.3
Peak mem (GB)
Hardware
Chip M3 Ultra (80c)
Memory 512 GB
GPU Cores 80
Software
oMLX v0.2.7
macOS macOS 15.6.1
Context 4,096
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 631.4 74.5 59.2 GB view
4k 693.6 68.2 59.3 GB current
8k 693.1 60.5 59.5 GB view
Batching Results
Batch Size TG tok/s Speedup
74.5 1.00×
108.8 1.46×
147.0 1.97×
187.6 2.52×