MSSQL pyodbc insert failing due to range of computer representation (8 bytes)

Can someone help me how to insert"OI-9786E543971246B29470AFA10414A1E2" value as a string to MSSQL db in python?

Error I get is:
"pyodbc.DataError: ('22003', "[22003] [Microsoft][ODBC SQL Server Driver][SQL Server]The floating point value '9786E' is out of the range of computer representation (8 bytes)."

        for client in clients:
            counts += 1
            insert_stmt += f"INSERT INTO [x].[dbo].[x] ([commcell],[clientName],[clientDisplayName],[clientHostName],[instanceId]) VALUES ('{client['commcell']}','{client['clientName']}','{client['clientDisplayName']}','{client['clientHostName']}','{client['instanceId']}') \r\n"
            if counts % 300 == 0:
                insert_stmt += "COMMIT TRANSACTION"
                cursor.execute(insert_stmt)
                cursor.commit()
                insert_stmt = "BEGIN TRANSACTION \r\n"
        if insert_stmt != "BEGIN TRANSACTION \r\n":
            insert_stmt += "COMMIT TRANSACTION"
            cursor.execute(insert_stmt)
            cursor.commit()

So basically if I would to remove OI- at the front of the value it would be ok..

Type of the value is string.
Tried column types in MSSQL to be as varchar(max), nvarchar(max), text and so on.

By concatenating strings this way you're causing a SQL injection attack against your own code.

The Insert a row in SQL Server's proof-of-concept tutorial shows how to safely pass values to any SQL query, not just INSERTs against SQL Server:

#Sample insert query
sql="""
INSERT INTO SalesLT.Product (Name, ProductNumber, StandardCost, ListPrice, SellStartDate) 
VALUES (?,?,?,?,?)"""

count = cursor.execute(sql,
'SQL Server Express New 20', 'SQLEXPRESS New 20', 0, 0, CURRENT_TIMESTAMP).rowcount
cnxn.commit()
print('Rows inserted: ' + str(count))

I modified the example a bit to make it clear that the SQL query doesn't include any values. This isn't a format string either. The parameter values never become part of the query itself. They're sent to the database server and used only after the query gets compiled into an execution plan.

At the very least, the python code should change to this:

sql = """
    INSERT INTO [x].[dbo].[x] ([commcell],[clientName],[clientDisplayName],[clientHostName],[instanceId]) "
    VALUES (?,?,?,?,?) 
"""

count = cursor.execute(sql,
    client['commcell'], 
    client['clientName'], 
    client['clientDisplayName'], 
    client['clientHostName'], 
    client['instanceId'])
...

It's also possible to insert multiple rows at a time using executemany and setting cursor.fast_executemany = True :

cursor.fast_executemany = True
..
tuples=[('foo','bar', 'ham'), ('hoo','far', 'bam')]
cursor.executemany(sql, tuples)
cursor.commit()

This caches all values in memory and sends all of them at once to the server