Vision Language Models (VLMs) are becoming increasingly important for document data extraction, but with the rapid growth of these models, it's essential to quickly evaluate the best options. VLMs integrate visual and textual information to understand and generate outputs based on multimodal inputs. They can be used for various applications such as visual question answering, image captioning, multimodal retrieval, and visual grounding. Some notable examples of VLMs include CLIP, LLaVA, Qwen2-VL-2B-Instruct, MiniCPM, Bunny, ChatGPT-4o-Mini, GPT4oMini, Claude 3.5, Gemini 1.5 Flash, and Gemini. To evaluate VLMs effectively, businesses should consider factors such as performance, scalability, and reliability. The results of this study show that Qwen is the best option for open-source models, while Gemini's free tier is the most cost-effective option for short-term predictions. However, it's essential to carefully evaluate prompts and perform error analysis to minimize hallucinations. Additionally, businesses should consider factors such as latency, compute resources, and accuracy when choosing a VLM. Ultimately, the choice of VLM depends on the specific business needs and requirements.