← Back to community benchmarks
Kimi-K2.6-Med-JANGTQ
Performance
4k
tokens
89.0
PP tok/s
19.1
TG tok/s
46021
TTFT (ms)
169.6
Peak mem (GB)
Hardware
Chip
M3 Ultra (80c)
Memory
512 GB
GPU Cores
80
Software
oMLX
v0.3.9.dev1
macOS
macOS 26.4.1
Context
4,096
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 90.2 | 19.4 | 167.5 GB | view |
| 4k | 89.0 | 19.1 | 169.6 GB | current |
| 32k | 73.5 | 14.7 | 185.6 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 19.4 | 1.00× |
| 2× | 29.7 | 1.53× |
| 4× | 40.7 | 2.10× |