There has been a rise in the proportion of England's teenagers getting the better GCSE or GNVQ grades - but less than the government's target increase.
The proportion getting no qualifications was 5.4%
New figures show 52.6% achieved five or more GCSEs at grades A* to C or their GNVQ equivalents. This was one point more than last year.
The target is for an average rise of two points a year from 2002 to 2006.
Specialist schools say they averaged 55.3% getting top grades, compared to 46.6% for the non-specialists.
Ministers want 92% of 16 year olds to have five or more qualifications at any grade, including English and mathematics, by 2004. This year, 86.3% have them.
The figure for the previous year - when the target was announced - was 87.1%, so there has been a decrease of 0.8.
Local education authorities fared better. The target was for 38% of the pupils in each area to achieve five or more grades A*-C at GCSE or GNVQ by 2004.
This year 138 authorities (92%) managed to do so, compared to 88% last time.
Girls continued to outperform boys, particularly at the higher grades.
Among girls, 57.8% got five or more grades A* to C, compared to 47.5% of boys.
The Education Minister Ivan Lewis said: "We have seen impressive strides in many schools that face challenging circumstances - for the second year running the poorest schools have improved significantly faster than the rest. They are narrowing the gap.
"These schools show what is possible and should be congratulated."
He said overall there was steady improvement but more needed to be done.
The shadow education secretary, Damian Green, commented on the fact that 38% of pupils got a C grade or above in English, maths and science.
"This is an alarming trend if we don't get the basics of education right then we will not get anything else right either," he said.
"It is terrifying that nearly two thirds of 16 year olds do not reach a reasonable standard in the three basic subjects."
At advanced level, for students aged 16 to 18 in schools and colleges, the average point score per A-level and AS-level entry was 76.9 - compared with 76.0 in last year's furore-hit results.
Analysis of the GCSE performance of specialist schools was carried out by Professor David Jesson of the University of York.
He said their relative performance over non-specialists of 8.67 points - nearly a fifth better - compared with a 7.4-point difference in 2002 "so the gap in performance is widening".
This was despite a large increase in the number of schools, up by 289 to 940 - against 1,989 non-specialists.
Prof Jesson told BBC News Online that the specialists did appear to be sustaining their advantage - but the acid test would be the "value added" performance data when it became widely available.
This compares not the raw results, but the way pupils improve during their time in a school.
The general secretary of the Secondary Heads Association, John Dunford, said heads remained disappointed that the main performance indicator was still the proportion gaining five A* to C grades.
"This figure does not give a true indication of a school's performance, which should be judged on value added and average points scores," he said.
The leader of the NUT teachers' union, Doug McAvoy, said: "It is no surprise that specialist schools are showing a higher increase than non-specialist schools given the additional resources thrown at them by government and the attention it focuses on them."
The Department for Education's national figures are collated from the provisional returns from all schools in England, which will form the basis of the performance tables at the end of the year.
They might well change. The department has compared this year's provisional figures with last year's final ones.
For example, it gives the figure for those getting five or more grades A* to C last year as 51.6 - but the same provisional figure at this time last year was 51.2 - so the rise appears to be bigger than the department has claimed.
The raw results released by the joint exam boards in August each year relate to all exam entries - not candidates. They do not show how many individuals took the exams.
These October statistics relate specifically to the performance of 15 year olds - the age of most students in England when they start their GCSE year. Official targets relate to 16 year olds - the age at the end of the academic year.