Open access journals in number theory

Here are a few thoughts:

• As several posters have already pointed out in the comments, many funding agencies are happy to accept an Arxiv preprint version as satisfying their "open access" requirement. This is certainly true for the big funders in the UK such as EPSRC, although you should check what your funder's rules are. (This is what's sometimes called "green open access".)

• Lots of publishers now offer the option called "gold open access". This is where you pay some extra fee at the time of publication to make your article freely available online. This is regarded with suspicion by many researchers, but that's another story. The point is that some institutions/funders have special pots of money available to pay these open-access fees, so it may end up costing you (and your research budget) little or nothing; it's worth looking into.

• A special case of the previous point is for journals that are published by your own funder. In this case the article-processing fees may be heavily discounted or waived entirely; if I remember correctly, this applies in the UK with the Royal Society, which allows holders of Royal Society grants to publish open-access in their journals without paying the usual fees.

That said, there are some excellent journals that are fully open-access and publish many number theory papers. "Documenta Mathematica" springs to mind, as does "LMS Journal of Computational Mathematics" for computational number theory.

The New York Journal of Mathematics (http://nyjm.albany.edu/) is a peer reviewed and free online general math journal that has published lots of papers in number theory. It has been publishing since the mid-1990s. It is not an ArXiv overlay journal. Published articles appear on the journal website as free downloadable PDF files. However, it is an electronic only journal, so if your funding requires a hard-copy print journal, it would not be allowed.