I learned Latin in a 6 week intensive course at a public university. I learned via a Reader's method (Latin Via Ovid) but using the traditional grammar-translation method, but based in an intensive graded reading course. That course was followed by a 6 week reading course reading Pliny. Intensive is the best way, way better than a slow drip method, I'm really happy I followed up the intensive course with an immersion in reading an original unfamiliar text (Pliny). I think if you can participate in an intensive course, you should do it. I can still decline many of the verbs, even though I do not use them continually. But, immersion (live language - ζωὴ Κοινή) is best, because you listen/read/speak many more repetitions than you would ever get reading 10-30 sentences.
It has been documented by modern psychology that the average person needs 30 to 70 repetitions to learn something (this could be driving a car through intersections, learning how to shift a stick-shift in a car with a manual transmission, learning the masculine dative 2nd declension singular form -ῳ, or cementing a word from working memory to permanent memory.) This means a student needs 70 reps of the masculine genitive 2nd declension singular, 70 of the masculine singular dative 2nd declension, ktl. How can students ever get this number of reps under their belt?
There are only two ways to get this kind of repetition: (1) Read, Read, Read. This is the only way for students who have been taught by the grammar-translation method to get enough reps to learn and internalize a form. It takes a lot of time to work one's way through a text when you have not internalized (via hearing) the morphological meaning of any given word or form. You either have to utilize a lot of graded readers or read a lot of texts and learn a lot of forms that you have to look up in a book, on a chart about 70-100 times. Or (2) participate in a spoken class where the average words heard in Greek per minute are 30+, (30x60 = 1800 per hour, probably more like 2500 per hour). I've calculated that the beginning classical grammar Athenaze has about 14,000 written words. Those who participate in a spoken class will use/hear/read 14,000 words in the first 7-10 weeks.
Unfortunately, most people taught by the grammar-translation method do not read enough texts to pass to a higher level of comprehension. And the constant grammar-translation method of always using the native language of the learner to describe the second language (2L = Koine) makes sure that the only way a grammar-translation student can understand a phrase is to translate it into his native language (in most people's cases, English). It's like telling a person that when they are driving a car, they have to think "look Left-Right-Left" at every intersection, rather than internalize that behavior. Or, to preach to them, "You will never understand the Greek, until you can accurately translate it in detail into English. Translating it is the only way you can understand it."
In the grammar-translation method, any connection to meaning to any given Greek word happens mentally, nor aurally. And hence, language acquisition of Koine does not occur. These students never start thinking Greek. They are taught to decode and think about Greek in terms of English language structure, word order, and idioms. Some of the quicker students are able to "catch on"; some very quick students can learn this way. They learn Greek like they learn math. A series of symbols and meanings.
But there are only about 20% of learners who can learn language this way. Students of ancient Greek want to learn how to "read with ease." It is really frustrating for a person, who has spent 60-90 hours in a class talking in English about a foreign language (Koine), and who have spent 200+ hours studying it, not to have immediate connections to the words and/or forms they are reading. Most at this point give up. Those few 20%'ers go on to a little more understanding, but even many of them get discouraged. In the end, perhaps 5% become somewhat adept at understanding a written text.
Students of Greek do not need to be SEALS, special OPS forces, for which qualified applicants are only .0005% of the population. It is not true that only a few select people will ever be able to read the GNT with ease. THE AVERAGE STUDENT CAN LEARN TO READ ANCIENT GREEK WITH THE SAME ABILITY THAT THEY COULD LEARN A MODERN LANGUAGE SUCH AS SPANISH. Even idiots [a technical term = IQ>80] who live in the English speaking world can speak English. What matters is HOW you learn the language. The goal of students of the spoken is that they want to read with fluency.
Immersion/Intensive learning is definitely the best way, but even then, it matters how you do it.