← Back to community benchmarks
Qwen3-Coder-Next
Performance
8k
tokens
993.4
PP tok/s
45.9
TG tok/s
8247
TTFT (ms)
43.4
Peak mem (GB)
Hardware
Chip
M2 Ultra (76c)
Memory
192 GB
GPU Cores
76
Software
oMLX
v0.2.6
macOS
macOS 15.0
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 802.0 | 50.3 | 43.0 GB | view |
| 4k | 972.6 | 47.5 | 43.2 GB | view |
| 8k | 993.4 | 45.9 | 43.4 GB | current |
| 16k | 972.5 | 41.7 | 43.9 GB | view |
| 32k | 891.1 | 30.6 | 44.7 GB | view |
| 64k | 729.6 | 20.9 | 46.5 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 50.3 | 1.00× |
| 2× | 91.9 | 1.83× |
| 4× | 149.1 | 2.96× |
| 8× | 213.3 | 4.24× |