The AMBOSS Score Predictor is a tool that predicts your exam score for USMLE® Step 1, USMLE® Step 2-CK, and USMLE® Step 3. You can enter your practice exam results (e.g., NBME® forms, AMBOSS or UWorld® Self-assessment results), and the dates on which you took these exams, and we'll predict your exam score on your target exam date. You’ll also receive a Predicted Score Range and the likelihood of passing your USMLE exam.
Our software offers the best accuracy in predicting scores when benchmarked against other popular score predictors.
How to use this tool:
- Start your prediction by clicking on “Share exams"
- Choose your target exam, i.e. the USMLE exam for which you would like to receive a score prediction
- Enter a target exam date
- Entering a target exam date is not mandatory for you to proceed but it is necessary for you to receive a score prediction
- Entering a target exam date is not mandatory for you to proceed but it is necessary for you to receive a score prediction
- Enter your practice exam results and the date on which you took the practice exam
- Only exams that are bolded in the form are relevant to the score prediction. Other exams will help us improve our prediction algorithm. None of these results will ever be shared with anyone.
- Only exams that are bolded in the form are relevant to the score prediction. Other exams will help us improve our prediction algorithm. None of these results will ever be shared with anyone.
- Submit your practice exam results by clicking on “Save exams”
- Get your Predicted Score, Predicted Score Range, and Likelihood of Passing
- You can see also see all the practice exams you’ve submitted
- You can see also see all the practice exams you’ve submitted
- If you want to share more practice exam results, click on “Edit exams”
- Click on “Save exams” to see your updated prediction
- Done with your USMLE exam and want to start studying for your next USMLE exam? Click on “Edit practice exams”, change your target exam, and start adding new practice exam results!
Relevant Practice Exams
Click here to see the practice exams we accept in order to calculate your predicted score and likelihood of passing.
Step 1:
- NBME® Comprehensive Basic Science Self-assessment Forms 25
- NBME® Comprehensive Basic Science Self-assessment Forms 26
- NBME® Comprehensive Basic Science Self-assessment Forms 27
- NBME® Comprehensive Basic Science Self-assessment Forms 28
- NBME® Comprehensive Basic Science Self-assessment Forms 29
- NBME® Comprehensive Basic Science Self-assessment Forms 30
- NBME® Comprehensive Basic Science Self-assessment Forms 31
- Free 120 that was last updated in June 2022
- AMBOSS Step 1 Self-assessment
- Uworld® Step 1 Self-assessment Form 1
- Uworld® Step 1 Self-assessment Form 2
Step 2:
- NBME® Comprehensive Clinical Science Self-assessment Form 9
- NBME® Comprehensive Clinical Science Self-assessment Form 10
- NBME® Comprehensive Clinical Science Self-assessment Form 11
- NBME® Comprehensive Clinical Science Self-assessment Form 12
- NBME® Comprehensive Clinical Science Self-assessment Form 13
- NBME® Comprehensive Clinical Science Self-assessment Form 14
- Free 120 that was last updated in July 2023
- AMBOSS Step 2 Self-assessment
- Uworld® Step 2 Self-assessment Form 2
- Uworld® Step 2 Self-assessment Form 1
Step 3:
- USMLE® Step 2-CK Score
- NBME® Comprehensive Clinical Medicine Self-assessment Form 5
- Free 137 that was last updated in Aug 2022
- Uworld® Step 3 Self-assessment Forms 1
- Uworld® Step 3 Self-assessment Forms 2
- High-yield computer-based case simulations
FAQs
How can I increase the accuracy of my prediction?
- Input proximity: The closer your practice exams are to your target exam date, the more accurate the Predicted Score and the smaller your Predicted Score Range.
- Number of inputs: More practice exams generally lead to a more accurate prediction and a smaller Predicted Score Range.
Why do I need to submit at least 2 relevant practice exam results to receive a prediction?
We cannot estimate your learning rate based on a single value. At least 2 practice exams are necessary to provide an accurate prediction. Adding more practice exam results will increase the accuracy of your prediction even further.
Why does one practice exam have to be within 4 weeks of my target exam?
Our model can predict USMLE scores accurately up to 4 weeks in the future. For dates that are further in the future (e.g., 6 weeks), predictions by our current models are less accurate. To ensure that your predictions are always accurate, we do not provide a prediction in these cases.
Why am I not getting a prediction?
There are a few reasons why you could not be getting a prediction:
1. You haven’t submitted a target exam date
- Enter a target exam date by clicking on “Edit exams.”
2. You’ve submitted less than two relevant practice exam results
- To provide an accurate prediction, we need at least 2 relevant practice exam results. Results that are relevant to the score prediction are bolded. You can also submit additional practice exam results (those that are not bolded). These results will not count towards your score prediction but will help us to improve our prediction algorithm and include the exams in the prediction in the future.
3. You’ve submitted less than one relevant practice exam result that is within 4 weeks of the target exam date.
- To be able to provide an accurate score prediction, one of your practice exam results should be within 4 weeks of your target exam date. Please submit one relevant practice exam result that is within 4 weeks of your target exam or change your target exam date so that it is within 4 weeks of your last submitted practice exam result.
How does the AMBOSS Score Predictor compare to the prediction made by the NBME forms?
The NBME CCSSA forms provide a 3-digit Predicted Score Range for your USMLE score if you were to take the exam within a week without gaining or losing any knowledge. The AMBOSS Score Predictor predicts your USMLE score for your target exam date, which can be up to 4 weeks in the future.
How exactly do the prediction algorithms work?
The prediction model for Step 1 and Step 2 uses a Mixed-effects model. This model considers the typical learning rate based on thousands of data points and your specific learning rate based on your individual information.
The prediction model for Step 3 uses the Joint normal distribution model. The scores of each input are assumed to follow a normal distribution and the most likely final score is calculated using Bayesian statistics.
Can I submit my USMLE Step exam score after I have taken the exam?
Yes! You can now submit your exam score by selecting the "Share final result" button on the Score Predictor page. Once on the Final result page, enter your exam, exam score, and exam date. Doing so will help us to improve our score predictor algorithm!