← Back to community benchmarks
Qwen3-Coder-Next
Performance
16k
tokens
952.3
PP tok/s
45.3
TG tok/s
17205
TTFT (ms)
62.4
Peak mem (GB)
Hardware
Chip
M2 Ultra (76c)
Memory
192 GB
GPU Cores
76
Software
oMLX
v0.2.10
macOS
macOS 26.3.1
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 648.6 | 56.4 | 61.6 GB | view |
| 4k | 920.3 | 54.6 | 61.7 GB | view |
| 8k | 981.0 | 51.1 | 62.0 GB | view |
| 16k | 952.3 | 45.3 | 62.4 GB | current |
| 32k | 877.7 | 35.3 | 63.3 GB | view |
| 64k | 729.5 | 22.5 | 65.1 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 56.4 | 1.00× |
| 2× | 93.7 | 1.66× |
| 4× | 142.7 | 2.53× |
| 8× | 195.4 | 3.46× |