← Back to community benchmarks
Qwen3-Coder-Next
Performance
16k
tokens
1,793
PP tok/s
50.8
TG tok/s
9137
TTFT (ms)
80.9
Peak mem (GB)
Hardware
Chip
M3 Ultra (80c)
Memory
512 GB
GPU Cores
80
Software
oMLX
v0.2.7
macOS
macOS 15.7.4
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,245 | 61.1 | 80.1 GB | view |
| 4k | 1,801 | 58.6 | 80.3 GB | view |
| 8k | 1,884 | 55.9 | 80.5 GB | view |
| 16k | 1,793 | 50.8 | 80.9 GB | current |
| 32k | 1,555 | 43.2 | 81.8 GB | view |
| 64k | 1,190 | 32.9 | 83.6 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 61.1 | 1.00× |
| 2× | 103.8 | 1.70× |
| 4× | 170.2 | 2.79× |
| 8× | 247.9 | 4.06× |