TEST #1298 RESULTS
08/03/2025 - 3:24 PM
ACCELERATOR
MODEL
23.8
tokens/s
generation
6.07
sec
time to first token
209
tokens/s
prompt
94
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M2 Pro 6P+4E
RAM
32GB
OS
Darwin
Kernel Release
25.0.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 25.0.0: Fri Jul 11 23:59:33 PDT 2025; root:xnu-12377.0.154.0.2~26/RELEASE_ARM64_T6020
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
252
tokens/s
27.3
tokens/s
4.10
sec
pp4096+tg256
235
tokens/s
19.6
tokens/s
17.46
sec
pp2048+tg256
250
tokens/s
23.8
tokens/s
8.22
sec
pp2048+tg768
213
tokens/s
20.4
tokens/s
9.68
sec
pp1024+tg1024
242
tokens/s
24.3
tokens/s
4.27
sec
pp1280+tg3072
149
tokens/s
20.4
tokens/s
8.67
sec
pp384+tg1152
240
tokens/s
25.5
tokens/s
1.64
sec
pp64+tg1024
210
tokens/s
27.0
tokens/s
340
ms
pp16+tg1536
87
tokens/s
25.8
tokens/s
219
ms