# Retrosynthetic Pathways \ **Choosing the retrosynthesis pathways from IBM RXN for Chemistry**\ First verify each SMILES and then select the 12class-tokens-2021-05-14 AI model version, and then select the Automatic Mode methodology and select High Quality. After that, only keep the retrosynthesis pathway predictions that have a confidence score >= 0.75.\ \ **The official IBM RXN for Chemistry platform guidance**\ The freely available forward prediction model and the single-step retrosynthetic model were trained using chemical reaction data automatically extracted from patents. The extraction process is not human-supervised and, although we made our best effort to reduce data noise, you should be aware that errors may still be present in the training data. Moreover, the distribution of chemical reaction data across the chemical reaction classes is not homogeneous, as some reaction classes are better represented than others. For instance, coupling reactions (i.e. Suzuki, Buchwald-Hartwig, etc.) appear more frequently than academic-type reactions (i.e. Diels-Alder, Michael additions, etc.), which are less represented. The presence of errors and the reaction class distribution issue in the training data may result in unreasonable or incorrect predictions for both models. Always use this prediction model judiciously with critical thinking. --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://retrosynthesis-ai.gitbook.io/retrosynthesis-ai/retrosynthesis-ai-platform/getting-started/retrosynthetic-pathways.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.