LLM Screener

It can be difficult to decide which Large Language Model (LLM) to use. There are many options (e.g. ChatGPT, Gemini, CoPilot, Claude, Perplexity etc) and most have several versions (e.g. ChatGPT 5, gpt5-mini and gpt5-nano). Additionally, they are typically optimized for different primary purposes (e.g. undertaking literature searches versus answering clinical questions), offer a variety of design features, and frequently undergo updates that change these features.

The LLM Screener was developed to provide a structured and credible process for people who are curious about what factors should be considered when choosing which LLM to use. Its design is based on regulatory, policy, ethical, and usability criteria derived from international and national literature published by reputable organizations such as the World Health Organization. It was tested with a spectrum of users and refined to meet the needs of healthcare researchers, clinicians and educators.

Use the LLM Screener

How to use the LLM Screener

1. Press the Get Started button below.

2. Enter your responses into each required field. You will need to answer these questions for each model that you are considering i.e. ChatGPT, Gemini and Claude.

  • If you need more information about the question, click on the small ‘I’ information item to see a more detailed description 
  • To find the answers to the questions about each model, you will typically need to find the most recent version of the background information on each model (i.e. parameters, training cut-off date etc.). It can be daunting to read through this information so one solution is to download this information for each model and then load these files into NotebookLM. You can then ask for the answer to these questions in NotebookLM to quickly get the responses that you can then enter into LLM Screener. 

3. Review the results

  • Scroll to the bottom of the screen to see the results. Compare the number of required criteria (the criteria that are mandatory/essential for you to be able to use the model) and the total criteria met (which includes the nonessential criteria) for each model that you are considering. You can download the results into a pdf if you wish. The model with the highest score is the model that would be best for you to use. 

You can download as a CSV file that can be used with Excel and Google Sheets.

The LLM Screener

If you have any questions about the LLM Screener, please contact Pascal Bolla at nexusdigitalhealth.com or message him on LinkedIn.

The Learning Technology Innovation Centre (LTIC) is located on the ancestral and unceded territory of the hən̓q̓əmin̓əm̓ – speaking xʷməθkʷəy̓əm (Musqueam) peoples. These lands have long been a place of learning for Musqueam, with cultural and traditional knowledge passed down from generation to generation for thousands of years.