← Back to community benchmarks
Nemotron-3-Nano-30B-A3B-TurboQuant
Performance
1k
tokens
1,398
PP tok/s
67.7
TG tok/s
733
TTFT (ms)
35.7
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
1,024
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,398 | 67.7 | 35.7 GB | current |
| 4k | 1,868 | 66.1 | 36.8 GB | view |
| 8k | 1,926 | 64.7 | 36.9 GB | view |
| 16k | 1,523 | 66.6 | 38.4 GB | view |
| 32k | 1,485 | 65.9 | 38.6 GB | view |
| 64k | 1,352 | 63.1 | 39.0 GB | view |
| 128k | 1,148 | 60.5 | 41.5 GB | view |
| 195k | 971.0 | 55.6 | 45.2 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 67.7 | 1.00× |
| 2× | 87.2 | 1.29× |
| 4× | 107.8 | 1.59× |
| 8× | 215.1 | 3.18× |