PeerJ, which will launch later this year, will also offer researchers in biomedicine an unlimited, lifetime right to publish for $259.
By paying either fee, scholars will become "members" of the journal.
Currently, academic authors who wish to make their papers open access are required to pay a fee for each article they publish. Elsevier's prestigious Cell titles charge $5,000 per paper, while even the giant open-access journal PLoS ONE charges $1,350.
PLoS ONE's former publisher, Peter Binfield, is one of the co-founders of PeerJ.
He agreed that the journal was being launched at an "auspicious" time, as discontent mounts over some publishers' alleged excess profits and resistance to open access.
A pledge to boycott Elsevier launched earlier this year has attracted 12,000 signatures from academics, while a petition asking the White House to adopt an open-access mandate for all publicly funded research passed its target of 25,000 signatories earlier this month.
The group of publishers, librarians and funders convened by David Willetts, the universities and science minister, to look at how best to manage a transition to open access for UK taxpayer-funded research is due to publish its conclusions next week.
But open-access advocates have expressed concern that the group, chaired by former Keele University vice-chancellor Dame Janet Finch, has indicated that it does not plan to comment on article fee levels.
Mike Taylor, an open-access advocate and a palaeontologist affiliated with the University of Bristol, said that the "audacity" of PeerJ's pricing had "changed the game permanently".
"In the early days of PLoS, conventional publishers sniggered at it and said it would never take off. They were very, very wrong. Now those same publishers will be looking very nervously at PeerJ," Dr Taylor said.
Dr Binfield said that PeerJ would emulate PLoS ONE's policy of accepting all methodologically sound papers, likely to be around 70 per cent of submissions. The journal's press materials say its low price is a result of having relatively low costs and its decision not to seek "excessive profit".
Dr Binfield estimated that biomedical papers have an average of around seven authors, all of whom - up to a maximum of 12 - would have to be members to publish in PeerJ.
For this reason, said Kent Anderson, editor-in-chief of the Scholarly Kitchen blog on academic publishing, PeerJ was unlikely in practice to work out much cheaper for researchers and was likely to be more administratively complex.
Dr Binfield said that PeerJ might force down other journals' article fees if it were successful, but added that this was a "rather grand ambition" for a start-up company.
He admitted that the influence of prestige on authors' publication decisions meant that journal publishing was "not a terribly price-sensitive market".
For this reason, PeerJ's principal selling point, in his view, would not be its price, but the fact that authors would be paying for membership rather than for one-off publication.