← Back to community benchmarks
Qwen3-32B-VL-Instruct-Polaris-GLM-Heretic-Uncensored-qx86-hi
Performance
1k
tokens
167.4
PP tok/s
10.6
TG tok/s
6119
TTFT (ms)
31.0
Peak mem (GB)
Hardware
Chip
M3 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.2.6
macOS
macOS 26.3.1
Context
1,024
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 167.4 | 10.6 | 31.0 GB | current |
| 4k | 135.7 | 9.4 | 31.8 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 10.6 | 1.00× |
| 2× | 8.5 | 0.80× |
| 4× | 10.2 | 0.96× |