← Back to community benchmarks
NVIDIA-Nemotron-3-Nano-30B-A3B
Performance
16k
tokens
4,927
PP tok/s
106.2
TG tok/s
3325
TTFT (ms)
21.0
Peak mem (GB)
Hardware
Chip
M5 Pro (20c)
Memory
64 GB
GPU Cores
20
Software
oMLX
v0.3.8
macOS
macOS 26.4.1
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,585 | 108.5 | 20.1 GB | view |
| 4k | 2,007 | 105.5 | 21.1 GB | view |
| 8k | 2,039 | 101.8 | 21.2 GB | view |
| 16k | 4,927 | 106.2 | 21.0 GB | current |
| 32k | 4,777 | 103.7 | 21.1 GB | view |
| 64k | 4,254 | 98.8 | 21.1 GB | view |
| 128k | 3,302 | 90.5 | 22.9 GB | view |
| 195k | 2,651 | 84.1 | 25.4 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 108.5 | 1.00× |
| 2× | 138.6 | 1.28× |
| 4× | 174.7 | 1.61× |
| 8× | 192.9 | 1.78× |