General Motors Corp. is taking its last chance at building small cars in the United States.

It is a $5 billion gamble called Saturn Corp. -- a newly created GM subsidiary that is expected to roll out its first subcompact model by 1990. It represents the industry's most important test for the balance of the decade.

GM, the nation's largest auto maker, virtually will stop producing small cars in this country if Saturn fails. But GM would continue to sell those models.

"You don't have to be a full-line manufacturer in order to be a full-line provider," GM Chairman Roger B. Smith told The Washington Post in a recent interview here. That means GM would fill its small-car needs with imports, mostly from Japan and Korea.

Ford Motor Co. and Chrysler Corp., GM's chief domestic rivals, would follow the leader -- escaping high U.S. production costs by bringing in high-quality small cars built overseas for far less money.

That development would wipe out several hundred thousand U.S. jobs. It also could leave U.S. car companies vulnerable to import attacks on their strongest end of the domestic auto market: mid-size and big cars.

But a successful launch of Saturn Corp. -- the creation of which was announced last week at the GM Technical Center in nearby Warren, Mich. -- could bring about major changes in the domestic auto industry. That, at least, is the thinking of most auto industry analysts and officials commenting on the project.

They cite three reasons for their belief:

* Saturn Corp. ultimately will be fashioned jointly by management and labor. Top officials of the United Auto Workers union are working closely with GM's top management to get the new company moving.

GM and the UAW have a history of contention -- one marked by strikes, costly management practices and protective union work rules, and triennial boosts in pay scales (now an average $23 hourly, including benefits) -- all of which have helped to place GM and other domestic auto makers at a disadvantage with foreign competitors.

Saturn Corp. has the leeway to work out a completely new labor agreement. Both the new company and the union are talking about different forms of compensation, different plant-floor procedures, and other changes to reduce production costs and ensure quality.

* Saturn Corp. will establish a new distribution system aimed at eliminating the negatives in car buying. High-pressure price negotiations and problems in new-car delivery to customers are the most frequently cited examples of those difficulties.

Saturn will have its own dealership network. Current GM dealers with high consumer satisfaction ratings are prime candidates for the new franchise.

* Of the $5 billion GM will spend on the Saturn project in the next 3 to 5 years, $3.5 billion will be invested in new plants and facilities. The rest will go into operating expenses.

Computerized, automated production and business administration systems will abound. Car assembly will involve a few major prefabricated modules and sections -- the entire engine and transmission-front-axle assembly could be one such section -- as opposed to the time-consuming, piece-by-piece method common in many U.S. auto plants today.

Saturn Corp. cars also will be "reverse designed" -- designed to take maximum advantage of production technology in the new assembly line. Currently it works the other way: The factory is set up to meet the engineering specifications of the car.

The use of modules also will make radical changes in the role of suppliers. Many suppliers now build and ship batches of one kind of part, such as a steering gear component. Under the modular system, suppliers could be required to build and ship whole sections of a car, which would be put together at the final assembly site.

A key objective of the proposed production changes is to reduce the number of worker-hours needed to make a single car. Currently, GM requires an average of 130worker-hours per car. GM wants to cut that to 70 or 75 hours per unit.

Fewer worker-hours means lower costs. For example, transferring 15 worker-hours to its suppliers -- either by going to a modular assembly system or using some other method -- would save GM up to $800 million annually in labor costs, according to a report by New York-based Sanford C. Bernstein & Co. Inc., a securities research firm.

But that potential saving compares only in-house and supplier-based labor costs in the United States, according to the Bernstein report, published last August. When compared with foreign labor costs, the possible savings that could be realized by shifting more production to suppliers "are much more significant," the Bernstein report said.

GM and its U.S. peers primarily are competing against Japan in the battle for the lucrative domestic small-car market, which accounted for 32 percent of all auto sales in this country in 1984. Mid-size or compact cars made up 46.7 percent of last year's sales. Although large and luxury cars had the biggest increases in rate of sales, they actually accounted for 21.3 percent of last year's U.S. car market, according to industry figures.

Imports, consisting mostly of small and mid-size cars, accounted for 23.4 percent of all cars sold in the United States last year. Most auto industry analysts agree that the import share would have been higher if Japanese auto shipments had not been restricted by quotas in place since April 1981.

For the Americans, the problem is this: Japanese auto makers build quality small cars at a unit price that is $1,500 to $2,000 less than comparable models built by their U.S. competitors.

"From a business standpoint, there is absolutely no reason for anyone to build a small or compact car in the United States if he can buy it somewhere else for $2,000 cheaper," said a knowledgeable U.S. auto industry official who requested anonymity for himself and his company.

"It just doesn't make any sense to do business that way. If American car companies can't find a way to get their costs down, and if the import gates open up, you can kiss domestic small-car production goodbye," the official said.

GM has been accused of doing just that with its much publicized "Asian strategy."

In its quest for a yearly supply of 1 million high-quality small cars, GM has entered a joint venture with Japan's Toyota Motor Corp. to produce up to 250,000 subcompacts a year in Fremont, Calif.

The venture, New United Motor Manufacturing Inc., actually is run by Toyota officials and will use 50 percent of Toyota componentry. GM also planned to import up to 300,000 small cars a year from its Japanese partners, Suzuki Motor Co. Ltd. and Isuzu Motors Ltd. But those import plans have been set awry by the quotas.

GM is looking for another substantial supply of small cars from its Korean partner, Daewoo Motor Co. GM's only homegrown small-car lines are the practically identical, 10-year-old Chevrolet Chevette and Pontiac 1000.

GM Chairman Smith repeatedly has denied that his company is abandoning small-car production in the United States. He says that he needs the imports and the joint-venture cars to fill in the huge gap in his small-car lineup until the Saturns start rolling.

But the Saturn has to be more than a car, Smith said. "If the Saturn is just another car that comes out of the wind tunnel, that might be exciting. But it won't be the lifesaver that it's got to be" if GM is going to continue producing small cars in the United States, he said.

James A. Mateyka, auto industry analyst with Booz Allen & Hamilton Inc., agrees. "If GM fails with Saturn, that'll certainly send a message to Ford and Chrysler that small cars can't be built here at competitive prices," Mateyka said.

"GM clearly is giving small-car production in this country one last go. And I guess that they've convinced themselves that going with a new corporation, working everything up from scratch, is the only way they can do it. I think Roger Smith wants to go down in history as the man who saved GM and the U.S. auto industry from the small-car hoards of the world," Mateyka said.