It will depend a bit on the cell number injected, how you harvest the cells and inject them: Recommend scraping the cells with a rubber policeman rather than trypsin and avoid using cells that were grown past confluency. We did titration of cell number versus initial tumor load for different cancer cells including MDA-MB-468,-231,-435 A549 etc and you get a good correlation between the initial tumor size and the time it takes to grow to a certain size relative to the cell number in the inoculum. Make sure, you aspirate the cells from the culture dish for their injection via a very wide-bore needle to avoid vacuum sheering of the cells. Switch to a narrow-bore needle to inject the mice to reduce pain as well as loss of cells at the injection site via the injection channel.
We published the titration and size of tumors in JNCI. I can send you a pdf if you can not download it. Reference: Zugmaier et al (1992) "Inhibition by pentosan polysulfate (PPS) of heparin-binding growth factors released from tumor cells and blockage by PPS of tumor growth in animals." J Natl. Cancer Inst. vol. 84 (22) pp. 1716-24
You should be able to get tumors with as low as 1 million of MDA-MB468 cells injected in a subcutaneous flanks or back of a nude mouse. You should not inject more than 5 million cells in order to have a steady growth of tumor and studies related to the progress. You have to mix the cells with matrigel (1:1 v/v) ratio before injecting so that cells could grow. I am a bit surprise to know your desire of getting a particular size of tumor and not sure if the studies will be with the tumor in mice or just to derive plenty of cells grown in mice, if that is the case then why not to pool a couple of tumors from different mice and generate the cells... just a suggestion.