If you want a price for 1k or 10k pcs, why ask here? None of us can possibly know what your distributor will quote you, but there's no harm at all in asking them for a quote. You'll get an idea of lead time and what the distributor or factory stock level is too.
If there's a compelling commercial case to switch to a Xilinx part instead, get the developer involved right now and explain the cost / benefit analysis you've done. Ask what the impact is likely to be of switching. FPGAs are anything but generic items for which drop-in replacements exist; you're asking the designer to learn a whole new tool chain and a whole new set of embedded logic blocks which may or may not do the job at all, and if they do, may or may not work the same way.
You may be indevertently throwing away a great deal of experience in specifying timing constraints, clock structures and relationships, or requirements for signal integrity... all manner of things which are only apparent to an experienced FPGA designer, and not at all obvious otherwise.
I use Altera exclusively, and whilst I'm sure competing products may be a better fit in some cases, the time and risk in switching have never been shown to me to be worthwhile. Maybe the day will come when I'm asked to do a cost sensitive, high volume FPGA-based design, but until then, low risk and short time-to-market have always been more important.
[Edited to fix heinous grammatical error which I blame on early morning coffee deficiency]