TEST #3687 RESULTS
06/05/2026 - 12:43 PM
ACCELERATOR
MODEL
53.3
tokens/s
generation
1.02
sec
time to first token
1248
tokens/s
prompt
403
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Core i7-4770 CPU @ 3.40GHz (haswell)
RAM
15.1GB
OS
Linux
Kernel Release
7.0.0-22-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #22-Ubuntu SMP PREEMPT_DYNAMIC Mon May 25 15:54:34 UTC 2026
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
1222
tokens/s
57.2
tokens/s
855
ms
pp4096+tg256
1241
tokens/s
45.3
tokens/s
3.32
sec
pp2048+tg256
1404
tokens/s
52.2
tokens/s
1.48
sec
pp2048+tg768
1404
tokens/s
51.0
tokens/s
1.48
sec
pp1024+tg1024
1484
tokens/s
54.3
tokens/s
707
ms
pp1280+tg3072
1443
tokens/s
48.8
tokens/s
905
ms
pp384+tg1152
1473
tokens/s
56.2
tokens/s
277
ms
pp64+tg1024
1194
tokens/s
58.0
tokens/s
69
ms
pp16+tg1536
367
tokens/s
56.9
tokens/s
59
ms