← Back to community benchmarks
Qwen2.5-Coder-1.5B-Instruct
Performance
4k
tokens
1,059
PP tok/s
29.6
TG tok/s
3870
TTFT (ms)
3.5
Peak mem (GB)
Hardware
Chip
M4 (10c)
Memory
16 GB
GPU Cores
10
Software
oMLX
v0.2.8
macOS
macOS 26.2
Context
4,096
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,043 | 31.2 | 3.3 GB | view |
| 4k | 1,059 | 29.6 | 3.5 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 31.2 | 1.00× |
| 2× | 40.3 | 1.29× |
| 4× | 78.6 | 2.52× |