← Back to community benchmarks
Qwen3.5-35B-A3B
Performance
8k
tokens
1,034
PP tok/s
56.3
TG tok/s
7926
TTFT (ms)
20.4
Peak mem (GB)
Hardware
Chip
M1 Ultra (64c)
Memory
128 GB
GPU Cores
64
Software
oMLX
v0.2.9
macOS
macOS 26.3.1
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,007 | 59.1 | 20.0 GB | view |
| 4k | 1,045 | 57.4 | 20.2 GB | view |
| 8k | 1,034 | 56.3 | 20.4 GB | current |
| 16k | 991.4 | 52.1 | 20.8 GB | view |
| 32k | 899.9 | 45.1 | 21.6 GB | view |
| 64k | 673.4 | 33.2 | 23.3 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 59.1 | 1.00× |
| 2× | 106.2 | 1.80× |
| 4× | 166.5 | 2.82× |
| 8× | 222.2 | 3.76× |