← Back to community benchmarks
gpt-oss-120b-MXFP4-Q8
Performance
32k
tokens
698.7
PP tok/s
30.0
TG tok/s
46899
TTFT (ms)
60.5
Peak mem (GB)
Hardware
Chip
M4 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.4
macOS
macOS 26.3.1
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 731.6 | 75.0 | 59.5 GB | view |
| 4k | 839.6 | 69.7 | 59.6 GB | view |
| 8k | 835.9 | 60.9 | 59.7 GB | view |
| 16k | 792.8 | 45.6 | 60.1 GB | view |
| 32k | 698.7 | 30.0 | 60.5 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 75.0 | 1.00× |
| 2× | 104.5 | 1.39× |
| 4× | 133.9 | 1.79× |
| 8× | 151.5 | 2.02× |