← Back to community benchmarks
gpt-oss-20b-MXFP4-Q4
Performance
4k
tokens
307.2
PP tok/s
37.4
TG tok/s
13333
TTFT (ms)
10.6
Peak mem (GB)
Hardware
Chip
M4 (10c)
Memory
32 GB
GPU Cores
10
Software
oMLX
v0.2.8
macOS
macOS 26.3.1
Context
4,096
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 302.5 | 40.5 | 10.6 GB | view |
| 4k | 307.2 | 37.4 | 10.6 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 40.5 | 1.00× |
| 2× | 47.0 | 1.16× |
| 4× | 54.2 | 1.34× |