TEST #4352 RESULTS

11/02/2026 - 9:16 PM

51.8

tokens/s

generation

876

ms

time to first token

1489

tokens/s

prompt

445

LocalScore

HOW YOU STACK UP
Explore All Results

Meta Llama 3.1 8B Instruct - Q4_K - Medium

SYSTEM
CPU
Intel Xeon CPU E5-2690 v4 @ 2.60GHz (broadwell)
RAM
62.8GB
OS
Linux
Kernel Release
7.2.13-100.fc43.x86_64
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #1 SMP PREEMPT_DYNAMIC Mon Oct 26 20:33:51 UTC 2026
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
1756
tokens/s
54.5
tokens/s
602
ms
pp4096+tg256
1400
tokens/s
45.9
tokens/s
2.95
sec
pp2048+tg256
1610
tokens/s
50.9
tokens/s
1.29
sec
pp2048+tg768
1603
tokens/s
49.9
tokens/s
1.30
sec
pp1024+tg1024
1732
tokens/s
52.1
tokens/s
610
ms
pp1280+tg3072
1676
tokens/s
48.5
tokens/s
783
ms
pp384+tg1152
1805
tokens/s
54.0
tokens/s
231
ms
pp64+tg1024
1442
tokens/s
55.3
tokens/s
62
ms
pp16+tg1536
378
tokens/s
54.7
tokens/s
60
ms