← Back to community benchmarks
Nemotron-Cascade-2-30B-A3B
Performance
32k
tokens
714.2
PP tok/s
54.2
TG tok/s
45881
TTFT (ms)
20.0
Peak mem (GB)
Hardware
Chip
M1 Ultra (48c)
Memory
64 GB
GPU Cores
48
Software
oMLX
v0.2.19
macOS
macOS 26.3.1
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 682.8 | 85.0 | 18.8 GB | view |
| 4k | 789.0 | 80.5 | 19.9 GB | view |
| 8k | 789.8 | 76.5 | 19.9 GB | view |
| 16k | 767.7 | 67.3 | 19.9 GB | view |
| 32k | 714.2 | 54.2 | 20.0 GB | current |
| 64k | 621.0 | 40.4 | 20.2 GB | view |
| 128k | 480.2 | 25.5 | 20.6 GB | view |
| 195k | 386.0 | 16.7 | 21.1 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 85.0 | 1.00× |
| 2× | 120.5 | 1.42× |
| 4× | 153.7 | 1.81× |
| 8× | 176.0 | 2.07× |