← Back to community benchmarks
gpt-oss-20b-MXFP4-Q8
Performance
64k
tokens
215.9
PP tok/s
11.0
TG tok/s
303520
TTFT (ms)
13.0
Peak mem (GB)
Hardware
Chip
M5 (10c)
Memory
32 GB
GPU Cores
10
Software
oMLX
v0.2.6
macOS
macOS 26.2
Context
65,536