← Back to community benchmarks

supergemma4-26b-abliterated-multimodal

M4 Max (40c) · 128 GB · 8bit · 2026-05-09
Performance
64k
tokens
1,011
PP tok/s
40.6
TG tok/s
64797
TTFT (ms)
30.0
Peak mem (GB)
Hardware
Chip M4 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.8
macOS macOS 15.7.4
Context 65,536
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,175 70.5 25.8 GB view
4k 1,394 69.9 26.3 GB view
8k 1,404 64.1 26.8 GB view
16k 1,350 64.0 27.1 GB view
32k 1,222 55.8 28.0 GB view
64k 1,011 40.6 30.0 GB current
Batching Results
Batch Size TG tok/s Speedup
70.5 1.00×
73.6 1.04×
123.7 1.75×