mschoi/project-euler-llm-benchmark
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

publish

Browse Source
This commit is contained in:
Michael Peter Christen
2025-02-19 06:11:35 +01:00
parent 062284068e
commit e2cb2e6c3c
2 changed files with 5 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
README.md
View File

@@ -37,6 +37,7 @@ The "Economic Score" is the average performance per bytes of model size (times 1
| qwen2.5:32b-instruct-q8_0 | | 25 | 32.8 | 8 | 32 | 9.73 | 10.22 | 5.91 | 2.98 |
| hf.co/bartowski/Qwen2.5-Coder-32B-Instruct-abliterated-GGUF:Q8_0 | | 25 | 32.8 | 8 | 32 | 11.58 | 8.69 | 6.27 | 1.62 |
| hf.co/bartowski/Qwen2.5-Coder-32B-Instruct-abliterated-GGUF:Q4_K_M | 16.40 | 46 | 32.8 | 4 | 32 | 9.55 | 9.3 | 5.71 | 1.94 |
| mistral-large:123b-instruct-2407-q4_K_M | | 12 | 122.6 | 4 | | 8.27 | 6.61 | | |
| GPT-4o-Mini | | | | 16 | 128 | 11.39 | 7.36 | 5.19 | 1.93 |
| deepseek-coder:33b-instruct-q4_K_M | | 44 | 33.0 | 4 | 16 | 9.55 | 10.72 | 0.0 | 3.03 |
| llama3.3:70b-instruct-q8_0 | | 10 | 70.6 | 8 | 128 | 8.93 | 8.06 | 4.29 | 3.17 |
@@ -48,6 +49,7 @@ The "Economic Score" is the average performance per bytes of model size (times 1
| hf.co/mradermacher/phi-4-abliterated-GGUF:Q4_K_M | 7.35 | 85 | 14.7 | 4 | 16 | 9.26 | 6.8 | 4.42 | 0.89 |
| phi4:14b | 7.35 | 84 | 14.7 | 4 | 16 | 9.64 | 6.91 | 3.14 | 0.84 |
| phi4:latest | 7.35 | 84 | 14.7 | 4 | 16 | 9.64 | 6.91 | 3.14 | 0.84 |
| hf.co/ozone-ai/0x-lite-Q4_K_M-GGUF:latest | | | 14.8 | | | 7.66 | 4.52 | | |
| phi4:14b-q8_0 | | 41 | 14.7 | 8 | 16 | 10.23 | 5.29 | 4.52 | 0.97 |
| hf.co/mradermacher/Llama-3.1-SauerkrautLM-70b-Instruct-GGUF:Q4_K_M | | 17 | 70.6 | 4 | 128 | 9.1 | 5.9 | 4.69 | 1.37 |
| vanilj/Phi-4:Q8_0 | | 38 | 14.7 | 8 | 16 | 9.06 | 5.73 | 3.52 | 0.84 |
@@ -126,17 +128,18 @@ The "Economic Score" is the average performance per bytes of model size (times 1
| yi:6b-chat-v1.5-q4_K_M | | 29 | 6.0 | 4 | 4 | 0.79 | 1.48 | 0.62 | 0.01 |
| codellama:34b-instruct-q4_K_M | | 5 | 34.0 | 4 | | 1.29 | 0.99 | 0.56 | 0.04 |
| qwen:32b-chat-v1.5-q4_K_M | | 5 | 33.0 | 4 | | 1.54 | 0.81 | 0.13 | 0.12 |
| llama3.2:latest | | 50 | 3.21 | 4 | 128 | 2.14 | 0.18 | 0.21 | 0.0 |
| qwen2.5:1.5b-instruct-q4_K_M | 0.75 | 101 | 1.5 | 4 | 128 | 1.94 | 0.26 | 0.15 | 0.0 |
| hf.co/bartowski/OpenThinker-7B-GGUF:Q4_K_M | | 19 | 7.62 | 4 | 32 | 0.88 | 0.77 | 0.53 | 0.62 |
| mixtral:8x7b-instruct-v0.1-q8_0 | | 2 | 46.7 | 8 | 32 | 1.44 | 0.65 | 0.23 | 0.0 |
| qwen:72b-chat-v1.5-q4_K_M | | 2 | 72.0 | 4 | 32 | 0.94 | 0.78 | 0.13 | 0.0 |
| command-r7b:7b-12-2024-q4_K_M | | 14 | 8.0 | 4 | 128 | 1.54 | 0.03 | 0.26 | 0.04 |
| hf.co/bartowski/agentica-org_DeepScaleR-1.5B-Preview-GGUF:Q4_K_M | | | 1.78 | | | 1.28 | 0.4 | 0.03 | 0.05 |
| olmo2:13b-1124-instruct-q4_K_M | | 8 | 13.7 | 4 | 4 | 1.59 | 0.06 | 0.01 | 0.03 |
| olmo2:7b-1124-instruct-q4_K_M | | 14 | 7.3 | 4 | 4 | 1.42 | 0.08 | 0.02 | 0.0 |
| qwen2.5:0.5b-instruct-q8_0 | 0.50 | 96 | 0.5 | 8 | 128 | 1.33 | 0.0 | 0.21 | 0.0 |
| codellama:13b-instruct-q4_K_M | | 7 | 13.0 | 4 | 16 | 0.52 | 0.84 | 0.05 | 0.08 |
| granite3.1-dense:2b-instruct-q8_0 | | 17 | 2.5 | 8 | 128 | 1.07 | 0.11 | 0.2 | 0.0 |
| llama3.2:latest | | 26 | 3.21 | 4 | 128 | 0.99 | 0.18 | 0.21 | 0.0 |
| smallthinker:3b-preview-q4_K_M | | 24 | 3.4 | 4 | 128 | 0.98 | 0.25 | 0.0 | 0.0 |
| mistral:7b-instruct-q4_K_M | | 12 | 7.0 | 4 | 32 | 0.48 | 0.71 | 0.04 | 0.0 |
| smallthinker:3b-preview-q8_0 | | 11 | 3.4 | 8 | 128 | 0.81 | 0.19 | 0.03 | 0.19 |

2
benchmark.json
View File

@@ -1012,7 +1012,7 @@
"_parameter_size": 1.78,
"clojure-100": 0.05,
"java-100": 0.4,
"python-100": 1.28
"python-100": 1.28,
"rust-100": 0.03
},
"granite3.1-dense:2b-instruct-q8_0": {