← Back to community benchmarks
Qwen3-Coder-Next
Performance
4k
tokens
1,304
PP tok/s
73.6
TG tok/s
3140
TTFT (ms)
62.5
Peak mem (GB)
Hardware
Chip
M4 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.6
macOS
macOS 26.4.1
Context
4,096
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,109 | 76.6 | 61.6 GB | view |
| 4k | 1,304 | 73.6 | 62.5 GB | current |
| 8k | 1,290 | 70.4 | 62.8 GB | view |
| 16k | 1,210 | 63.6 | 63.5 GB | view |
| 32k | 1,057 | 52.4 | 64.9 GB | view |
| 64k | 831.2 | 39.0 | 67.7 GB | view |
| 128k | 560.0 | 18.8 | 73.3 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 76.6 | 1.00× |
| 2× | 113.8 | 1.49× |
| 4× | 153.0 | 2.00× |
| 8× | 196.5 | 2.57× |