← Back to community benchmarks
Unsloth-Qwen3.6-35b-a3b
Performance
1k
tokens
910.6
PP tok/s
51.9
TG tok/s
1125
TTFT (ms)
35.4
Peak mem (GB)
Hardware
Chip
M2 Ultra (60c)
Memory
64 GB
GPU Cores
60
Software
oMLX
v0.4.2rc1
macOS
macOS 15.7.7
Context
1,024
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 910.6 | 51.9 | 35.4 GB | current |
| 4k | 1,059 | 52.2 | 36.2 GB | view |
| 8k | 1,075 | 45.0 | 36.5 GB | view |
| 16k | 1,034 | 43.0 | 37.1 GB | view |
| 32k | 939.6 | 40.4 | 38.5 GB | view |
| 64k | 645.2 | 30.7 | 40.1 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 51.9 | 1.00× |
| 2× | 55.4 | 1.07× |
| 4× | 117.6 | 2.27× |