← Back to community benchmarks
Qwen3.5-VL-4B-8bitCRACK
Performance
64k
tokens
251.1
PP tok/s
8.7
TG tok/s
260978
TTFT (ms)
9.9
Peak mem (GB)
Hardware
Chip
M4 (10c)
Memory
16 GB
GPU Cores
10
Software
oMLX
v0.2.9
macOS
macOS 26.2
Context
65,536
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 396.3 | 21.6 | 5.8 GB | view |
| 4k | 401.1 | 20.9 | 6.0 GB | view |
| 8k | 389.6 | 19.0 | 6.3 GB | view |
| 16k | 365.8 | 16.2 | 6.8 GB | view |
| 32k | 323.7 | 12.3 | 7.8 GB | view |
| 64k | 251.1 | 8.7 | 9.9 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 21.6 | 1.00× |
| 2× | 43.1 | 2.00× |
| 4× | 64.9 | 3.00× |
| 8× | 67.5 | 3.12× |