Dear Editors, Please find enclosed our manuscript entitled “Knowledge Graph Extraction in a Practical Context” which we are submitting for consideration for publication in the special issue of Neurosymbolic Artificial Intelligence: Special Issue on Knowledge Graphs and Neurosymbolic AI. This manuscript is an extension of the paper "From Text to Knowledge Graph: Comparing Relation Extraction Methods in a Practical Context" by Roos M. Bakker and Daan L. Di Scala, part of the proceedings of the workshop GeNeSy’24: First International Workshop on Generative Neuro-Symbolic AI, co-located with ESWC 2024, May 26, 2024, Hersonissos, Greece. In our manuscript, we address the challenge of extracting knowledge graphs and ontologies from text. We present a comparative analysis of relation extraction methods for knowledge graph extraction. The methods are assessed within a real-life scenario, aiming for a full graph with quality comparable to manually developed graphs. Previous methodologies often relied on automatically extracted datasets and a limited range of relation types, consequently constraining graph expressivity. Moreover, these datasets typically feature short or simplified sentences, failing to capture the complexity of real-world texts like news messages or research papers. The results show that GPT models outperform other relation extraction methods in quantitative metrics. However, qualitative analysis reveals that alternative approaches like REBEL and KnowGL excel in leveraging external world knowledge to enrich the graph beyond textual content alone. This highlights the importance of considering methods that not only extract relations directly from text but also incorporate supplementary knowledge sources to enhance the richness and depth of resulting knowledge graphs. This manuscript is highly relevant for researchers in the neurosymbolic AI domain, where integrating symbolic knowledge into neural AI architectures is crucial. Knowledge graphs and ontologies play a significant role in this integration, bridging the gap between symbolic reasoning and neural networks. Creating and maintaining these structures remains a long-standing challenge in ontology learning. By leveraging neural AI architectures to automatically extract knowledge graphs and ontologies from text, we aim to reduce the burden on developers and enable automatic, objective extensions. This approach illustrates the principles of neurosymbolic AI by combining the strengths of neural processing with symbolic representation, improving the overall capability and interpretability of AI systems. The extensions on the previous paper are as follows: 1. Extended Related Work on all topics, especially on Ontology Learning, 2. Extended our experiments and results to include the most state-of-the-art GPT model, GPT-4 omni. It's high performance in graph extraction further influences our results and conclusions. 3. Extended the results and analysis of them from both the new method and the older methods, and 4. Additional discussion points on graph and annotation aspects. Our extended paper is not currently under consideration by another journal. Included in this submission is the PDF of the manuscript. The LaTeX files can also be provided if desired. Thank you for your consideration and time, we look forward to the reviews. Kind regards, Roos Bakker