← Back to community benchmarks
Mistral-Small-3.2-24B-Instruct-2506
Performance
8k
tokens
365.2
PP tok/s
38.4
TG tok/s
22433
TTFT (ms)
14.6
Peak mem (GB)
Hardware
Chip
M3 Ultra (60c)
Memory
256 GB
GPU Cores
60
Software
oMLX
v0.3.4
macOS
macOS 26.4
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 365.5 | 47.4 | 13.1 GB | view |
| 4k | 374.1 | 43.0 | 13.9 GB | view |
| 8k | 365.2 | 38.4 | 14.6 GB | current |
| 16k | 345.9 | 30.3 | 15.7 GB | view |
| 32k | 311.2 | 23.4 | 18.3 GB | view |
| 64k | 260.3 | 14.8 | 23.3 GB | view |
| 128k | 195.3 | 8.1 | 33.8 GB | view |
| 195k | 152.6 | 5.4 | 44.7 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 47.4 | 1.00× |
| 2× | 70.9 | 1.50× |
| 4× | 80.1 | 1.69× |
| 8× | 92.0 | 1.94× |