← Back to community benchmarks
Qwen3-Coder-Next
Performance
1k
tokens
817.5
PP tok/s
47.4
TG tok/s
1253
TTFT (ms)
43.0
Peak mem (GB)
Hardware
Chip
M2 Ultra (76c)
Memory
192 GB
GPU Cores
76
Software
oMLX
v0.2.6
macOS
macOS 15.0
Context
1,024
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 817.5 | 47.4 | 43.0 GB | current |
| 4k | 975.5 | 51.0 | 43.2 GB | view |
| 8k | 992.8 | 47.1 | 43.4 GB | view |
| 16k | 972.6 | 40.5 | 43.9 GB | view |
| 32k | 891.9 | 30.7 | 44.7 GB | view |
| 64k | 728.7 | 21.1 | 46.5 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 47.4 | 1.00× |
| 2× | 93.1 | 1.96× |
| 4× | 149.3 | 3.15× |
| 8× | 216.3 | 4.56× |