TEST #28 RESULTS
03/27/2025 - 11:50 PM
ACCELERATOR
MODEL
63.1
tokens/s
generation
650
ms
time to first token
1959
tokens/s
prompt
575
LocalScore
HOW YOU STACK UP
Explore All ResultsMistral Nemo Instruct 2407 - Q4_K - Medium
SYSTEM
CPU
AMD EPYC 7H12 64-Core Processor (znver2)
RAM
503.5GB
OS
Linux
Kernel Release
6.8.0-52-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #53~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Wed Jan 15 19:18:46 UTC 2
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
2300
tokens/s
65.5
tokens/s
461
ms
pp4096+tg256
1936
tokens/s
56.8
tokens/s
2.13
sec
pp2048+tg256
2157
tokens/s
62.4
tokens/s
966
ms
pp2048+tg768
2172
tokens/s
61.5
tokens/s
959
ms
pp1024+tg1024
2300
tokens/s
64.1
tokens/s
460
ms
pp1280+tg3072
2229
tokens/s
60.3
tokens/s
589
ms
pp384+tg1152
2257
tokens/s
65.0
tokens/s
185
ms
pp64+tg1024
1789
tokens/s
66.2
tokens/s
50
ms
pp16+tg1536
493
tokens/s
65.7
tokens/s
46
ms