Check below input/output data, and the employee third highest employee salary.

Input data:

empid       ename      salary
----------- ---------- -----------
1           pqr        25000
2           abc        25000
3           xyz        37000
4           pqr        35200
5           ymf        40000
6           sfs        40000
7           wer        37000
8           cvb        37800

Require output:

empid       ename      salary
----------- ---------- -----------
3           xyz        37000
7           wer        37000

Create a table with sample data:

create table tbl_employee 
(
	empid int primary key
	,ename varchar(10)
	,salary int	
)
GO

insert into tbl_employee values 
(1,'pqr',25000)
,(2,'abc',25000)
,(3,'xyz',37000)
,(4,'pqr',35200)
,(5,'ymf',40000)
,(6,'sfs',40000)
,(7,'wer',37000)
,(8,'cvb',37800)
GO

Solution 1 (Correlated Subquery):

select * from tbl_employee as a
where 3 = (select count(distinct salary) from tbl_employee as b where a.salary<=b.salary)

Solution 2 (DENSE_RANK()):

with cteEmp as
(
	select
		empid
		,ename
		,salary
		,DENSE_RANK() over (order by salary desc) as rnk
	from tbl_employee
)
select 
	empid
	,ename
	,salary
from cteEmp
where rnk = 3

Solution 3 (MAX() & CROSS APPLY):

SELECT a.ename, b.salary
FROM 
(
	SELECT MAX(salary) Salary
    FROM tbl_employee d
    WHERE d.Salary !=
	(
		SELECT MAX(salary) Salary
		FROM tbl_employee e 
		WHERE e.Salary != (SELECT MAX(salary) FROM tbl_employee)
	)
	and d.Salary != (SELECT MAX(salary) FROM tbl_employee)
) b
CROSS APPLY (SELECT ename FROM tbl_employee WHERE salary = b.salary ) a