Neither my memory nor the Wikipedia history suggest any formal IBM role in the development of 'Windows' itself, at least not through Windows 3.11. It wasn't a joint project (though I suppose the closeness of the partnership likely meant source/fix/feature sharing behind the scenes).
Then, IBM was formally cooperating with Microsoft on something they thought would be OS/2. To the extent anything co-developed then wound up in Windows NT, that was never IBM's intent. Again, Windows was the Microsoft-controlled branch, not the partnership product.
That would roughly match the assertion by rational-future about which you were dismissive.