TEST #1014 RESULTS

06/07/2025 - 1:43 AM

25.8

tokens/s

generation

330

ms

time to first token

4014

tokens/s

prompt

679

LocalScore

HOW YOU STACK UP
Explore All Results

Meta Llama 3.1 8B Instruct - Q4_K - Medium

03570105140tokens/s (higher is better)NVIDIA RTX PRO6000 BlackwellWorkstationEditionNVIDIA RTX 6000Ada GenerationThis System(NVIDIA GeForceRTX 4090)Apple M2 Pro6P+4E+16GPUApple M3 Pro5P+6E+14GPUIntel Corei7-8550U CPU @1.80GHz(skylake)Intel PentiumCPU G4600 @3.60GHz(skylake)
SYSTEM
CPU
AMD Ryzen 7 7800X3D 8-Core Processor (znver4)
RAM
63.6GB
OS
Windows
Kernel Release
10.0
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
4819
tokens/s
25.9
tokens/s
254
ms
pp4096+tg256
4353
tokens/s
19.4
tokens/s
998
ms
pp2048+tg256
4680
tokens/s
26.9
tokens/s
485
ms
pp2048+tg768
5766
tokens/s
23.8
tokens/s
391
ms
pp1024+tg1024
5100
tokens/s
27.6
tokens/s
240
ms
pp1280+tg3072
4970
tokens/s
24.7
tokens/s
292
ms
pp384+tg1152
5062
tokens/s
28.1
tokens/s
115
ms
pp64+tg1024
1119
tokens/s
28.0
tokens/s
94
ms
pp16+tg1536
261
tokens/s
27.5
tokens/s
106
ms