← Back to community benchmarks

gpt-oss-120b-heretic-v2-mxfp4-q8-hi

M3 Ultra (60c) · 256 GB · 4bit · 2026-03-18
Performance
8k
tokens
1,047
PP tok/s
65.4
TG tok/s
7823
TTFT (ms)
59.8
Peak mem (GB)
Hardware
Chip M3 Ultra (60c)
Memory 256 GB
GPU Cores 60
Software
oMLX v0.2.17
macOS macOS 26.2
Context 8,192
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 871.4 68.9 59.5 GB view
4k 1,054 71.7 59.6 GB view
8k 1,047 65.4 59.8 GB current
16k 982.1 55.8 60.0 GB view
32k 855.6 46.3 60.6 GB view
Batching Results
Batch Size TG tok/s Speedup
68.9 1.00×
92.6 1.34×
126.9 1.84×
163.4 2.37×