← Back to community benchmarks
MiniMax-M2.5
Performance
8k
tokens
675.4
PP tok/s
30.1
TG tok/s
12130
TTFT (ms)
175.9
Peak mem (GB)
Hardware
Chip
M3 Ultra (80c)
Memory
512 GB
GPU Cores
80
Software
oMLX
v0.2.6
macOS
macOS 26.3
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 631.0 | 39.8 | 174.0 GB | view |
| 4k | 725.3 | 34.8 | 175.0 GB | view |
| 8k | 675.4 | 30.1 | 175.9 GB | current |
| 16k | 575.8 | 23.7 | 177.8 GB | view |
| 32k | 437.7 | 16.2 | 181.7 GB | view |
| 64k | 294.1 | 10.0 | 189.6 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 39.8 | 1.00× |
| 2× | 57.5 | 1.44× |
| 4× | 73.7 | 1.85× |
| 8× | 104.3 | 2.62× |