← Back to community benchmarks
NVIDIA-Nemotron-3-Nano-30B-A3B
Performance
4k
tokens
1,864
PP tok/s
66.5
TG tok/s
2198
TTFT (ms)
44.9
Peak mem (GB)
Hardware
Chip
M5 Pro (20c)
Memory
64 GB
GPU Cores
20
Software
oMLX
v0.3.6
macOS
macOS 26.4.1
Context
4,096
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,363 | 67.7 | 43.8 GB | view |
| 4k | 1,864 | 66.5 | 44.9 GB | current |
| 8k | 1,910 | 64.3 | 45.0 GB | view |
| 16k | 1,989 | 66.7 | 44.8 GB | view |
| 32k | 1,846 | 66.0 | 45.7 GB | view |
| 64k | 1,498 | 63.8 | 49.7 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 67.7 | 1.00× |
| 2× | 92.9 | 1.37× |
| 4× | 113.9 | 1.68× |
| 8× | 127.8 | 1.89× |