TEST #1877 RESULTS
11/28/2025 - 12:59 AM
ACCELERATOR
MODEL
29.3
tokens/s
generation
4.04
sec
time to first token
288
tokens/s
prompt
128
LocalScore
HOW YOU STACK UP
Explore All ResultsQwen3-Coder-30B-A3B-Instruct-1M - Q4_K - Medium
SYSTEM
CPU
Intel Xeon CPU E5-2670 0 @ 2.60GHz (sandybridge)
RAM
377.8GB
OS
Linux
Kernel Release
5.15.0-161-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #171-Ubuntu SMP Sat Oct 11 08:17:01 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.3
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
351
tokens/s
30.3
tokens/s
2.95
sec
pp4096+tg256
324
tokens/s
21.4
tokens/s
12.69
sec
pp2048+tg256
340
tokens/s
27.1
tokens/s
6.06
sec
pp2048+tg768
346
tokens/s
27.2
tokens/s
5.95
sec
pp1024+tg1024
351
tokens/s
31.1
tokens/s
2.95
sec
pp1280+tg3072
334
tokens/s
25.7
tokens/s
3.86
sec
pp384+tg1152
321
tokens/s
34.1
tokens/s
1.22
sec
pp64+tg1024
157
tokens/s
34.6
tokens/s
433
ms
pp16+tg1536
69
tokens/s
32.5
tokens/s
256
ms