← Back to community benchmarks
Qwen3-Coder-Next
Performance
16k
tokens
783.5
PP tok/s
42.6
TG tok/s
20911
TTFT (ms)
80.9
Peak mem (GB)
Hardware
Chip
M2 Ultra (60c)
Memory
128 GB
GPU Cores
60
Software
oMLX
v0.2.6
macOS
macOS 26.3.1
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 676.0 | 51.2 | 80.1 GB | view |
| 4k | 806.7 | 54.5 | 80.3 GB | view |
| 8k | 825.1 | 52.0 | 80.5 GB | view |
| 16k | 783.5 | 42.6 | 80.9 GB | current |
| 32k | 719.6 | 36.5 | 81.8 GB | view |
| 64k | 560.5 | 21.4 | 83.6 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 51.2 | 1.00× |
| 2× | 89.7 | 1.75× |
| 4× | 135.6 | 2.65× |
| 8× | 207.6 | 4.05× |