I abandoned giving midyear and final exams to my third-year Latin students about five years ago.  Now, as my discipuli were soon to learn to their dismay, that certainly did not mean they were being given a free period in which to nap, snack, or enjoy some quality loitering.  Instead, I had tossed the tests so that I could assess what my students had learned by means of the translation-based project whose instructions and rubric I have included below.
The exams which I used to give as my major summative assessments were more effective, I suspected, in showing what specifically students knew, rather than whether they were able to relevantly employ that knowledge.  If I had already bombarded them with a semester of quizzes and tests to evaluate their grasp of the vocabulary and the grammar and the culture, why not conclude with an evaluation of their ability to utilize all those skills in concert?  The comprehension and translation portions of the exams, to be fair, did accomplish this goal to an extent, and I did consider just making that section the entire assessment.  A desire to help my kids develop practical skills, however, nudged me toward making this a project.  There are plenty of times when a standard test is absolutely appropriate, but, at the end of the day, for the majority of our lives, we do not need to demonstrate our knowledge of a subject in sudden, short bursts.  Using what we have learned to independently complete some project over an extended period of time, on the other hand, is a skill that is critical to succeeding in any profession one might pursue.
The instructions for the project are, I hope, adequately lucid.  Each student is assigned a different passage of roughly equivalent length from an author.  The length of the passage depends upon the difficulty of the author and the amount of time (usually at least three weeks) I have given them to work on it.  At a time they have chosen during the week of midyears/finals, they meet with me one-on-one to discuss their specific passage and the rest of their project.  The intent of this friendly interrogation is not only to evaluate their skills as a whole, but also to make certain that their work is their own — that they did not simply memorize a translation they found elsewhere, or dump the text into Google Translate, or ask another student to complete portions of the project for them.
Lastly, as is obvious, my version of this project is pretty grammar-translationey.  Nevertheless, I think it should not be too difficult to adapt it to whichever pedagogical approach or authors you favor.  And I would love to hear any recommendations which you might have regarding how I might improve it!
 
The Project
L3H Final Project Rubric