metadata
base_model:
- Qwen/Qwen2.5-Coder-7B-Instruct
tags:
- ServiceNow
Automated benchmark (due to time constrains)
This benchmark compares the Qwen2.5-Coder-7B-Instruct model with this servicenow finetune, Qwen QwQ 32B and Quasar-Alpha (Secret new model on Openrouter, revealed as a Pre-Release of GPT 4.1, coding comparable a bit better than DeepSeek V3, https://openrouter.ai/openrouter/quasar-alpha, https://openrouter.ai/openai/gpt-4.1). DeepSeek R1 evaluated the results of each benchmark question.
Please note: This process definitly needs some improvements, for a general overview it should be good enough tho
Results were okay but not as good as i wanted, definitly taking another look at the training data and different approaches