← Back to community benchmarks
Qwen3-Coder-Next
Performance
8k
tokens
1,034
PP tok/s
47.6
TG tok/s
7927
TTFT (ms)
80.5
Peak mem (GB)
Hardware
Chip
M3 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.2.8
macOS
macOS 15.7.3
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 856.0 | 48.7 | 80.1 GB | view |
| 4k | 1,042 | 44.4 | 80.3 GB | view |
| 8k | 1,034 | 47.6 | 80.5 GB | current |
| 16k | 973.3 | 40.2 | 80.9 GB | view |
| 32k | 814.9 | 28.2 | 81.8 GB | view |
| 64k | 589.3 | 17.9 | 83.6 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 48.7 | 1.00× |
| 2× | 68.8 | 1.41× |
| 4× | 86.0 | 1.77× |
| 8× | 104.1 | 2.14× |