← Back to community benchmarks
Qwen3.5-35B-A3B
Performance
1k
tokens
502.9
PP tok/s
67.5
TG tok/s
2036
TTFT (ms)
19.3
Peak mem (GB)
Hardware
Chip
M1 Max (32c)
Memory
64 GB
GPU Cores
32
Software
oMLX
v0.3.5
macOS
macOS 26.1
Context
1,024
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 502.9 | 67.5 | 19.3 GB | current |
| 4k | 595.0 | 63.9 | 20.0 GB | view |
| 8k | 600.8 | 58.6 | 20.4 GB | view |
| 16k | 577.2 | 50.6 | 21.0 GB | view |
| 32k | 525.4 | 42.1 | 22.3 GB | view |
| 64k | 422.5 | 27.5 | 25.0 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 67.5 | 1.00× |
| 2× | 95.7 | 1.42× |
| 4× | 117.4 | 1.74× |