The Working Group on Expanding Access, chaired by former Keele University vice-chancellor Dame Janet Finch, is basing its projections of what a wholesale switch to "gold" - or "author-pays" - open access would cost the UK academy on a "cost-neutral" fee of £1,450 per article published.
However, according to the minutes of the group's penultimate meeting in April, members noted that the figure was "no more than an average" that would vary according to discipline and various other factors such as take-up rates both within the UK and abroad.
For these reasons, they "agreed strongly that it would not be appropriate" to set a "benchmark" figure for article fees.
The group, which includes funders, librarians and publishers, was convened by universities and science minister David Willetts and its report, expected later this month, is likely to form the basis of government policy.
Discussions are said to have been robust, but Times Higher Education understands that the final meeting in May saw the members reach broad agreement on the main issues.
However, the decision not to endorse any guideline article fee will disappoint those who advocate a move to open access partly as a means to moderate what they regard as commercial publishers' excessive profit margins.
Tim Gowers, the University of Cambridge professor of mathematics whose pledge in January to boycott Elsevier has so far been echoed by 12,000 academics, told THE that he was concerned that the Finch commission was not intending to address the issue of publishers' profits.
"There may be good reasons for not specifying a benchmark article fee, but I see no argument against at least establishing a principle that such fees should be for the purpose of covering publication costs rather than replacing lost profits," Professor Gowers said.
"Taxpayers' money that should be spent on research is being spent on publishers' very large profits, and the danger with the gold open-access model is that it will not deal with this problem."
At the very least, he said, the group should specify that "if two journals in the same area are of comparable quality, then funding bodies will cover publication charges at the level of the cheaper one".
Robert Kiley, head of digital services at the Wellcome Trust and a Finch group member, said the existence of downward pressure on article fees was demonstrated by the fact that fees charged by recent start-ups were considerably less than the "standard" charge of $3,000 (£1,900).
But Peter Murray-Rust, reader in molecular informatics at Cambridge, said there was little market pressure on publishers to bring down costs or improve their products.
"There is even less market force in the gold model, where publishers can charge what they like with no regulation," he said.
"The market often resembles personal vanity products, where only the brand matters and cost of production is irrelevant."
He said that even a £1,500 fee would represent a "seriously overpriced market".