← Back to community benchmarks
openclaw-qwen3.5-2b-classifier
Performance
16k
tokens
4,283
PP tok/s
210.8
TG tok/s
3826
TTFT (ms)
2.4
Peak mem (GB)
Hardware
Chip
M3 Ultra (60c)
Memory
96 GB
GPU Cores
60
Software
oMLX
v0.2.6
macOS
macOS 26.3.1
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 2,057 | 272.1 | 2.0 GB | view |
| 4k | 3,870 | 262.1 | 2.2 GB | view |
| 8k | 4,254 | 243.5 | 2.2 GB | view |
| 16k | 4,283 | 210.8 | 2.4 GB | current |
| 32k | 3,989 | 167.8 | 3.0 GB | view |
| 64k | 3,347 | 118.6 | 4.5 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 272.1 | 1.00× |
| 2× | 449.5 | 1.65× |
| 4× | 606.5 | 2.23× |
| 8× | 782.5 | 2.88× |