Langfuse's September 2023 update introduces several key features and improvements aimed at enhancing the evaluation and integration of language model applications. Notably, the update includes model-based evaluations via a Python SDK, allowing users to run evaluation functions on production data and score results within Langfuse. The introduction of Datasets in beta enables systematic testing of LLM applications by managing collections of inputs and expected outputs. In-app analytics have been enhanced with a new query engine, facilitating the development of dashboards for deeper insights. New integrations with Flowise, Langflow, and LiteLLM expand the toolkit for users, while complex filtering for tables and the ability to share traces via public links improve data accessibility and collaboration. Users can export data directly from the Langfuse UI in various formats for tasks such as fine-tuning. These updates aim to streamline the evaluation, analysis, and integration processes for users working with large language models.