← Back to community benchmarks
Qwen3.6-40B-Claude-4.6-Opus-Deckard-Heretic-Uncensored-Thinking
Performance
1k
tokens
502.7
PP tok/s
12.4
TG tok/s
2037
TTFT (ms)
41.0
Peak mem (GB)
Hardware
Chip
M5 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.8
macOS
macOS 26.5
Context
1,024
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 502.7 | 12.4 | 41.0 GB | current |
| 4k | 563.0 | 12.3 | 43.0 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 12.4 | 1.00× |
| 2× | 23.9 | 1.93× |
| 4× | 33.5 | 2.70× |