TEST #2273 RESULTS
01/13/2026 - 3:01 AM
ACCELERATOR
MODEL
44.1
tokens/s
generation
691
ms
time to first token
1995
tokens/s
prompt
503
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Xeon W-2133 CPU @ 3.60GHz (skylake-avx512)
RAM
94.3GB
OS
Linux
Kernel Release
6.8.0-90-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #91-Ubuntu SMP PREEMPT_DYNAMIC Tue Nov 18 14:14:30 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
2326
tokens/s
45.4
tokens/s
463
ms
pp4096+tg256
1722
tokens/s
41.0
tokens/s
2.40
sec
pp2048+tg256
2068
tokens/s
43.7
tokens/s
1.01
sec
pp2048+tg768
2080
tokens/s
43.3
tokens/s
1.01
sec
pp1024+tg1024
2360
tokens/s
44.7
tokens/s
456
ms
pp1280+tg3072
2236
tokens/s
42.5
tokens/s
595
ms
pp384+tg1152
2675
tokens/s
45.2
tokens/s
166
ms
pp64+tg1024
2090
tokens/s
45.7
tokens/s
53
ms
pp16+tg1536
401
tokens/s
45.4
tokens/s
61
ms