Benchmark of ChatMDR/ChatIVDR models
ChatMDR and ChatIVDR Free users can choose from two models – Fast and Smart – and Plus and Unlimited Users also get the Reasoning option, which breaks down a challenge into multiple thought steps and comes to a more detailed and thought-through answer.
But how does each model perform in a head-to-head comparison? We tested this by asking the model the same, MDR-related question. Where needed, we continued to prompt a little more to get to an answer we could use. Please click on each of the three links below to see how the models performed.
The question we asked each model was:
I have a new class IIa device that is part of a sampling group. What is the procedure to bring it to the market as quickly as possible?
Fast model response to question at ChatMDR 3.1 release
Fast Model response to class IIa sampling group question The Fast model starts describing a self-certification process, but fails to mention the involvement of the notified body. Only after additional questioning (so…
Smart model response to question at ChatMDR 3.1 release
Smart Model response to class IIa sampling group question The Smart model correctly mentions notified body involvement, but needs to be prompted again to come up with the smart use of a…
Reasoning model response to question at ChatMDR 3.1 release
Reasoning Model response to class IIa sampling group question The Reasoning model immediately understands what we are trying to achieve and provides a clear overview, with references and quoted pieces of text…