MINNEAPOLIS (AP) _ A drop in the price of Cray Research Inc. stock following the company's announcement that it is ending a multimillion-dollar supercomputer project should not suggest long-term problems, the company says.

Cray announced Wednesday that Senior Vice President Steve Chen, who directed the supercomputer project, plans to leave Cray to produce the project on his own.

The news sent Cray's stock tumbling. It closed at $104.50 on the New York Stock Exchange Wednesday, down $8.50 from Tuesday's close.

''We expected that the immediate perception would be 'Oh, no, you lost one of your top designers,''' said Cray spokeswoman Tina Bonetti. ''It will take awhile for people to realize there's more to it than that.''

Sewall Hodges, a security analyst for Arnhold & S. Bleichroeder of New York, said some stockholders may be overreacting.

''If you're going to sell the stock, you shouldn't do it because of this,'' he said. ''But any slice of bad news, or news that is interpreted as bad, is all it takes'' to make the stock slide.

Cray Chairman John Rollwagen said Chen's supercomputer project ''has grown significantly beyond our original vision, both in terms of technological risk and budget.''

''We believe it no longer meets the objectives and style of Cray Research,'' Rollwagen said. ''With this announcement we are refocusing on the fundamentals that have made Cray Research an industry leader.''

About half of the 180 employees working on the project at Cray's Chippewa Falls, Wis., plant will be laid off, Bonetti said.

Chen will leave the company in the near future and may take some employees involved with the project with him in his effort to continue on his own, she said.

Chen joined Cray as chief design engineer in 1979 and was later promoted to vice president of development, where he was responsible for directing the development of the company's X-MP line of supercomputers.

Supercomputers are distinguished from the biggest mainframe computers by virtue of their brute strength and are mainly used for science and engineering. Capable of more than a billion arithmetic calculations a second, the machines can simulate three-dimensional situations such as automobile crashes and airplane flight.

This year Cray spent $16 million to $17 million on the project that began in late 1985, Bonetti said, adding that she didn't know how much the company spent on the project in 1985 and 1986.

She said $3 million to $4 million from the project will be written off as a loss during the third quarter, but the project won't have any effect on Cray's revenues after 1987.

Chen's project, which was still in the design stages and was internally called the MP project, focused on both hardware and software design, with a high-level parallel processing architecture, Bonetti said. Chen had envisioned marketing a supercompeter that used 64 processors by 1991, Bonetti said. Cray now markets supercomputers with four processors.

Research showed that the project was too unwieldly, Bonetti said. The project was orginally expected to cost about $50 million, but recent projections showed it would cost closer to $100 million, she said.

She said the project also would have included research into technology that is not used in supercomputer manufacturing today, which the company believes is too risky at this time.

''I think it's obviously a business decision, and Chen doesn't have his name on the door of the company,'' said Hodges.

Hodges said Chen and Seymour Cray, who founded the company in 1972, were working on similar supercomputers independent of each other, and Cray won out.

''I don't see (the decision to drop Chen's project) to have much impact on revenues; it may be even better,'' Hodges said. ''They'll have less research expenses because they won't have dual research for one project. There will be a little less waste. Some people will leave.''

Bonetti denied that the idea that competition between Cray, who designed the company's Cray line, and Chen, who designed the X-MP line, was the reason for the decison.

She said Cray decided it did not want to risk adding the MP, which would have amounted to a third line for the company.

But she said the company will retain some elements of Chen's project that will be helpful to some of its other development projects.