← Back to community benchmarks

gpt-oss-20b

M3 Max (40c) · 48 GB · unknown · 2026-03-12
Performance
1k
tokens
737.8
PP tok/s
74.2
TG tok/s
1388
TTFT (ms)
13.1
Peak mem (GB)
Hardware
Chip M3 Max (40c)
Memory 48 GB
GPU Cores 40
Software
oMLX v0.2.8
macOS macOS 26.3.1
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 737.8 74.2 13.1 GB current
4k 901.3 67.3 13.1 GB view
Batching Results
Batch Size TG tok/s Speedup
74.2 1.00×
72.2 0.97×
112.0 1.51×