← Back to community benchmarks
GLM-54.8bit
Performance
32k
tokens
58.0
PP tok/s
10.7
TG tok/s
565396
TTFT (ms)
442.7
Peak mem (GB)
Hardware
Chip
M3 Ultra (80c)
Memory
512 GB
GPU Cores
80
Software
oMLX
v0.2.9
macOS
macOS 15.4.1
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 178.6 | 16.1 | 420.0 GB | view |
| 4k | 178.9 | 13.4 | 422.3 GB | view |
| 8k | 153.2 | 12.9 | 424.9 GB | view |
| 16k | 116.9 | 11.9 | 430.9 GB | view |
| 32k | 58.0 | 10.7 | 442.7 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 16.1 | 1.00× |
| 2× | 22.7 | 1.41× |
| 4× | 28.9 | 1.80× |