← Back to community benchmarks
Qwen3-Coder-Next
Performance
64k
tokens
1,298
PP tok/s
50.8
TG tok/s
50476
TTFT (ms)
45.4
Peak mem (GB)
Hardware
Chip
M3 Ultra (80c)
Memory
512 GB
GPU Cores
80
Software
oMLX
v0.5.1
macOS
macOS 26.5.1
Context
65,536
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,889 | 68.2 | 43.1 GB | view |
| 4k | 2,186 | 67.2 | 43.8 GB | view |
| 64k | 1,298 | 50.8 | 45.4 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 68.2 | 1.00× |
| 2× | 119.4 | 1.75× |
| 4× | 183.5 | 2.69× |
| 8× | 279.7 | 4.10× |