← Back to community benchmarks

gpt-oss-20b-MXFP4-Q8

M4 (10c) · 24 GB · 4bit · 2026-04-13
Performance
1k
tokens
260.8
PP tok/s
33.1
TG tok/s
3926
TTFT (ms)
11.7
Peak mem (GB)
Hardware
Chip M4 (10c)
Memory 24 GB
GPU Cores 10
Software
oMLX v0.3.4
macOS macOS 26.4
Context 1,024
Batching Results
Batch Size TG tok/s Speedup
1× 33.1 1.00×
2× 43.2 1.31×
4× 41.5 1.25×
8× 27.1 0.82×