TEST #1176 RESULTS
07/04/2025 - 10:31 AM
ACCELERATOR
MODEL
4.9
tokens/s
generation
72.11
sec
time to first token
19
tokens/s
prompt
11
LocalScore
HOW YOU STACK UP
Explore All ResultsDeepseek-R1-0528-Qwen3-8B - Q4_K - Medium
SYSTEM
CPU
Intel Xeon CPU E5-2620 v3 @ 2.40GHz (haswell)
RAM
755.8GB
OS
Linux
Kernel Release
6.8.0-58-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #60~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Fri Mar 28 16:09:21 UTC 2
RUNTIME
Name
llamafile
Version
0.9.3
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
19
tokens/s
5.9
tokens/s
53.01
sec
pp4096+tg256
18
tokens/s
4.6
tokens/s
233.49
sec
pp2048+tg256
19
tokens/s
4.6
tokens/s
110.11
sec
pp2048+tg768
19
tokens/s
4.6
tokens/s
109.70
sec
pp1024+tg1024
20
tokens/s
4.8
tokens/s
52.62
sec
pp1280+tg3072
19
tokens/s
4.6
tokens/s
66.68
sec
pp384+tg1152
20
tokens/s
5.1
tokens/s
19.04
sec
pp64+tg1024
21
tokens/s
5.2
tokens/s
3.27
sec
pp16+tg1536
19
tokens/s
5.1
tokens/s
1.05
sec