EARLY NEXT MONTH a panel of Pentagon officials will make one of those multi-billion-dollar decisions that help keep the American defense budget in the fiscal stratosphere. In all likelihood, the panel will make the most expensive choice, and the evidence is compelling that it will be the wrong choice as well.

The weapon is the Maverick, a so-called "smart" weapon which contains its own elaborate guidance system. It is intended to be used by Air Force fighters to destroy Warsaw Pact tanks in any future Soviet invasion of western Europe. The Air Force wants to order 61,000 Mavericks at a cost currently estimated at $5 billion.

Maverick is the embodiment of the pervasive, relentless passion in the Pentagon for highly complex, costly "solutions" to problems for which advanced but simpler and less expensive alternatives are available. In this case the alternative is a remarkable cannon of proven reliability as a tank killer.

In the view of many critics of Pentagon purchasing policies, the available cannon is a fine weapon, while the Maverick is a turkey. These experts argue that unless the Maverick can be proven to be a reliable improvement on the existing weapon, the Pentagon should not buy it.

The Pentagon likes the cannon, and is buying it in large quanitities. But it also wants the Maverick -- even at a cost of $5 billion or more.

The Air Force wants the Pentagon panel meeting next month to authorize "limited pilot production" of 200 second-generation Mavericks, a foot in the door that the Air Force hopes will lock in the ultimate purchase of 61,000 missiles. Alton V. Keel, assistant secretary of the Air Force for research, development and logistics, assured the House Armed Services Committee last March, "This missile will work and we can produce it."

Such arguments didn't impress committee staff expert Anthony R. Battista, who has studied Maverick test videotapes, interviewed test pilots, and flown in the rear seat of a test aircraft. "What justification have you to keep this program alive?" he asked. "It is five years now and we are getting the same arguments in favor of keeping it alive now as we did then . . ." The Maverick, he told Keel, "would probably get more pilots killed than they would kill targets."

The General Accounting Office wants the panel to hold off, saying that five years of operational testing of the new Maverick have failed to show that it "can be used effectively by U.S. military personnel in combat." The missile did poorly even under "very favorable test conditions," such as alerting pilots to "what to look for," the GAO stressed in a report released July 2.

Ten days later, a reporter learned, two Mavericks, the last of 10 tested since February, flunked an operational test in Utah. The previous eight were found to have a less than 30 percent probability of working properly after being airborne for 14 hours, though they were supposed to be 85 percent reliable. After repairs, however, six of the eight missiles reportedly destroyed their targets, while two more failed. The Air Force declined to comment on either the GAO report or the recent testing.

Still, the odds are strong that the Pentagon panel -- which rarely kills any weapons program that has advanced as far as Maverick has -- will approve the first purchase of missiles.

The chairman of the panel, Under Secretary of Defense Richard D. DeLauer, is already on record as a staunch supporter of Maverick. "I've looked at the test data so far," he said in an interview early this year. "I've looked at the design. One of the guys who designed it sat in that chair. I have high confidence in his technical ability and his honesty to me." He was speaking of one of his predecessors as director of Pentagon research and engineering, Malcolm R. (Mal) Currie, vice president for missiles of Hughes Aircraft Co., Maverick's manufacturer.

"He and I went over the whole goddamn design," DeLauer said. "And, you know, we talked what the hell is the problem, how can we get (a component) to work a little better, what the production problems are. You know, we sat there as a couple of engineers. That's my goddamn life."

Moreover, the Pentagon claims that there is no alternative to the new Maverick. If the DeLauer panel turns it down, "I frankly don't know where we'd turn," Maj. Gen. James H. Marshall, the Air Force Chief of development and production said in a December interview. "There's no other solution hanging on the horizon to go grab."

Yet even as he spoke, Air Force A10 squadrons were armed and poised for combat with the most lethal tank-killer available to tactical aircraft anywhere in the world. The weapon is a simple, inexpensive cannon -- the first in 30 years that the Air Force has developed for airplanes.

The gun uses a radically new 30-millimeter shell with an unique incendiary penetrator. Nearly twice as heavy as lead, and moving at a speed of 3,240 feet a second, it cuts easily through side or rear armor. The heat generated in tearing through thick steel makes the penetrator burn fiercely, and bits and pieces scatter around the tank compartment, igniting fires and exploding ammunition. The rate of fire -- from seven barrels -- is 4,200 rounds per minute. Usually, a strafing burst of a mere 30 rounds or so kills a tank.

In operational tests at Nellis Air Force Base, Nev., in 1975 and 1976, "passes" by A10s against the sides or rears of 10 Soviet T62 main battle tanks gutted and destroyed six and totally immobilized the other four. The 60 percent kill rate was four times higher than had been predicted by computer studies during development of the gun, called the GAU8/A.

Cheerleaders for the new cannon are found even among advocates of the missile. "I'll never downgrade it," says Maverick's program requirements manager, Maj. Grant G. (Nick) Nicolai. A top jet pilot, he has logged more than 1,000 hours in the A10, which carries two missiles under each wing and the GAU8/A gun in its nose. For tank-killing at short-range, he says, the gun is "excellent . . . the best thing in the world for it's purpose."

However, Nicolai and other avocates of the Maverick think "its purpose" is too limited. They want another tank killer that can be used from much greater distances; they want a tank killer that is a missile, not a gun. As Maj. Nicolai put it, he thinks a pilot would want to have both weapons. "As a fighter pilot," he said, "I don't want to get my ass shot down. I want to live to come back to drink beer at the bar."

Others disagree. Col. John M. Verdi, a retired Marine Corps aviation weapons expert, led one of the first Marine fighter squadrons in Vietnam combat. For him, the choice between Maverick and the gun is no choice at all.

"If the question is, which would you rather take in against the target at interest, I would unhestitatingly say, 'guns,' " said Verdi, who holds the Legion of Merit and 30 air medals. "I wouldn't want that other sonofabitch," he told a reporter from his home in Santa Ana, Calif. "You have to have a gun on a plane that's going to fight somebody. Anything else is absurd."

The Air Force pays about $20 each for the 30mm armor-piercing rounds, which are made of "depleted" uranium of very low radioactivity. A 30-round burst costs $600 -- less than one-hundredth the estimated cost of a single advanced Maverick.

The Air Force has been buying the "DU" rounds, which were independently developed and produced by Aerojet Ordnance Co. and Honeywell Corp.for about seven years. Now large amounts of the ammunition are "pre- positioned," mainly in Europe, for possible use in battle. By the end of 1983, the Air Force expects to have an adequate war reserve of tens of millions of rounds.

Not for many years after that, if ever, could there be enough of the new Mavericks. It would take a decade s of Hafter 1984, when full production may begin, for Hughes Aircraft, and possibly a second manufacturer, to build the desired 61,000, and the Air Force does not so much as claim that this number would constitute an adequate war reserve.

One crucial advantage of the cannon is logistics, and it has been powerfully documented by Col. Robert G. Dilger, a retired fighter pilot who managed the GAU8/A gun and DU ammo programs for the Air Force.

For a squadron of 16 functioning A10s, he said in a May 1980 talk to the American Defense Preparedness Association, a six-man ammunition ground crew is needed to reload the GAU8/A guns, each with 1,174 rounds; a 70-man crew is needed to reload Mavericks, four on each airplane. He also figured that the supplies airlifted in by a single C130 enable the squadron to make 520 firing passes with the gun, but only 60 with Mavericks.

To reload a single aircraft, the gun crew needs a half hour, sometimes as little as 15 minutes, compared with about 75 minutes needed by the Maverick crew, Dilger said. For each pass against a tank, the approximate ground time needed for the gun is two minutes; for the missile, 18 minutes. The more time an aircraft spends on the ground, the fewer the combat sorties it can fly.

A former Pentagon specialist in air warfare says he would forget Maverick and substitute under each wing of the A10 a four-barrel gun enclosed in a pod, the GEPOD 30, so as to enhance the aircraft's anti-tank capability while increasing its ability to take out air defenses. The pod and the GAU8/A fire the same armor-piercing ammunition. Both guns were developed by the General Electric Co. working with the Air Force.

Notably, James P. Wade Jr., Under Secretary DeLauer's principal deputy, acclaims the 30mm pod. In a letter to the GAO, he pointed out that it evolved "from existing, proven hardware," that it has the "capability to kill multiple vehicles in a single pass," that its accuracy permits it to be used "in close proximity to friendly troops," and that it can be carried on numerous tactical aircraft in addition to the A10 -- the F4, F5, F15, F16 and A7.

The Tactical Air Force "does not have one single weapon which has the flexibility to attack many different types of targets," including tanks, armored personnel carriers, supply depots and bunkers, Wade wrote. The 30mm pod "provides a practical near-term solution" because it "has the flexibility and lethality against this wide spectrum of targets and has demonstrated kill potential against enemy armor arrays."

Yet the Air Force refused for three straight years to fund the GEPOD 30, until Congress forced the money on it for fiscal 1980. This was but one of many signs that the enduring Air Force romance is not with the cannon, but with successive generations of Maverick missiles.

"I know they (Air Force leaders) don't like the gun," says Thomas S. Hahn, a member of the House Armed Services Committee staff from 1975 to 1980. "Missilery is what people are all enthused about."

"They hated guns, because guns are far too simple and are likely to compete with glamorous high technology," says a former Pentagon weapons analyst. "The Maverick backers were saying the gun (generically) is obsolete, even though we haven't begun to scratch the available new gun technology."

The romance with Maverick originated in the early 1960s, when President Kennedy revived sagging interest in conventional tactical weapons by abandoning the policy of massive retaliation for "a wider choice than humiliation or all-out nuclear action." The Air Force found itself with no effective way to attack small armored targets. Its thin-skinned jet aircraft, built for the entirely different mission of dropping bombs, were catastrophically vulnerable to defensive automatic weapons fire. Its 20mm Vulcan cannon, excellent for aerial combat, pelted a tank as ineffectually as raindrops pelt a window.

This situation spawned a fervent wish for a "stand-off" air-to-ground weapon, meaning one that would be fired, in theory, from outside the lethal envelope of air defenses. In July 1964 the wish became the father to an Air Force "operational requirement" for what would become the original Maverick, a daytime-only, television-guided missile that works on contrasts between a target and its lighter or darker immediate surroundings.

The Air Force "tested" the TV Maverick in ideally stark conditions in the sunswept Nevada and New Mexico deserts, firing it at dark tanks squatting on white sand. Despite warnings from critics that the missile would not work against normally camlouflaged targets, the Air Force spent about $500 million to buy 20,000 Mavericks and sent some into combat in Southeast Asia. There, green enemy tanks melded into green vegetation, making the missiles nearly useless, according to Gen. John W. Vogt, who commanded U.S. Air Forces in the region.

Meanwhile, the Air Force stockpiled TV Mavericks as its primary antitank weapon in Europe, where the environment is green, wet and overcast much of the time. In 1975, Vogt, who had become commander of U.S. Air Forces in Europe, ordered the first realistic operational test, and the missile failed dismally.

This did not catch the Air Force flatfooted: it had prepared for just such a contingency by starting to develop the second-generation, heat-seeking Maverick, which uses infrared rays to sense thermal contrasts between an object and its warmer or cooler close surroundings.Its glory is that darkness does not affect it, so it can "see" 24 hours a day.

But the new Maverick has two main weaknesses. First, on the tiny cockpit screen, the images of many objects, such as sun-warmed rocks and burning bushes, can resemble the images of tanks. Second, it can be defeated by simple humidity; by thermal clutter, such as hot shell holes and flaming battlefield wreckage, and by enemy countermeasures, such as tank-gun firings. For the Maverick, says Battista, the House Armed Services staffer, all of this makes for a "world down there (that) is very, very cruel."

The Air Force objection to the cannon is twofold: it can be used only in daylight, and then only with excessive danger, because the aircraft cannot stand off a great distance from the target.

"The argument against the gun as the primary antiarmor weapon is that the pilot can't get in close enough to use it before getting shot," says Col. Verdi. The error in that is that standoff only buys the first (wrong) shot; then you have to close (in on) the enemy."

Some analysts go so far as to assert that the missile has no real utility at all in combat -- even in daytime -- and one of them says that this is shown by a fundamental paradox:

To spot a camouflaged object shrouded in battlefield haze and make sure it's a tank, an enemy tank, a pilot must see it with his eyes. At "eyeball" range -- no more than 3,000 feet -- he can't fire the missile, because it needs more distance to overcome the perturbations of launch so that its guidance system can take over. But at the much greater stand-off ranges at which the missile could home accurately, he can't spot the target out of the cockpit, and he can't be certain that the object outlined on his five-inch-square cockpit screen is the enemy tank.

A simple operational test would have revealed the dilemma even before the first TV Maverick was bought, says C.E. Myers Jr., a former Pentagon director of air warfare.

By persevering in its overriding commitment to Maverick, the Air Force has created a situation full of seeming anomalies:

It developed and has acclaimed the GAU8/A gun and the unique ammunition, but embraces only avericks -- even the TV version it stopped buying in 1978.

As a primary argument for the new Maverick, the Air Force and the Pentagon have emphasized for many years that its infrared sensor works as well in darkness as in daylight, giving it night and adverse-weather capability. But they have yet to spend any money on adapting an infrared viewer as a nighttime gunsight. The Fairchild Republic Co., with its own money, did such an adaptation, and, it says, the experiment showed the cannon to be as accurate in darkness as in daylight.

Summing up the arguments: the rigorously tested gun is much safer to use than Maverick. It is cheap and classically simple, and its shells go where they are aimed. The questionably tested new Maverick is unreliably workable in combat. It is highly dangerous to use, costly and complex. It frequently fails to go where it is aimed, because its infrared guidance can be deceived and defeated. And the logistics strongly favor the gun.