AI Model Review is dedicated to providing transparent reviews and information about generative AI models.
We are currently in the midst of an AI explosion with new generative AI models being released frequently, and by a wide range of organizations, from government institutions, large dominant tech companies to individual creators.
For consumers like us, it can be daunting to understand what the different models have to offer.
While there are information sources about model performance and model rankings , we aim to provide a more qualitative perspective, by asking the models to respond to a standardized set of questions. This way we can directly compare how the models respond and get an intuition about the strengths and weaknesses of different models.
There are currently over 75 models available to compare, and we will continue to add new models as they become available. With the goal of including as many models as possible. Our comparisons include both: models that can only be accessed via a service such as ChatGPT and models that can be downloaded to run locally like Llama. In the model comparison cards we provide links that will take you to the model page where you can download the weights for the locally available models.
We take our standard set of questions and prompt the models 4 times for each question. Each question is prompted with a different temperature setting. We want to see how the temperature setting affects the model’s response. You can change the temperature by selecting from the dropdown at the bottom of the model comparison card to see the different responses.
The AI Model Review Newsletter is the best way to stay informed and get updates on:
You may get slightly different results when you try these prompts yourself, this may be because:
We are not AI researchers, AI developers or affiliated with any organizations which develop AI technologies. We are avid enthusiasts and consumers of AI technology and believe these tools will empower all of us to reach our full potential in productivity and creativity.
We provide model responses from all of our tests openly on our model comparison page so you do not need to rely on trust alone to get a good understanding of the capabilities of AI models. While our reviews are written from our own perspective, our perspective is that of the average consumer. Those who share in this perspective may find value in our reviews.
We welcome comments, suggestions and feedback. Please feel free to contact us via email at AIMR@AIModelReview.com