TEST #23 RESULTS
03/27/2025 - 10:21 PM
ACCELERATOR
MODEL
308
tokens/s
generation
106
ms
time to first token
12344
tokens/s
prompt
3294
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Xeon CPU E5-2660 v3 @ 2.60GHz (haswell)
RAM
62.8GB
OS
Linux
Kernel Release
5.15.0-130-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #140~20.04.1-Ubuntu SMP Wed Dec 18 21:35:34 UTC 2024
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
16516
tokens/s
337
tokens/s
65
ms
pp4096+tg256
11770
tokens/s
262
tokens/s
352
ms
pp2048+tg256
13756
tokens/s
307
tokens/s
152
ms
pp2048+tg768
12917
tokens/s
295
tokens/s
162
ms
pp1024+tg1024
15344
tokens/s
309
tokens/s
70
ms
pp1280+tg3072
14112
tokens/s
283
tokens/s
94
ms
pp384+tg1152
17269
tokens/s
319
tokens/s
25
ms
pp64+tg1024
8731
tokens/s
331
tokens/s
10
ms
pp16+tg1536
684
tokens/s
325
tokens/s
26
ms