This post sets our three GIT model papers against the real GIT 2026 Paper I, question by question, and gives the full score with every miss named. We wrote the three papers before the exam on 24 October 2026. Many model papers say they "predict" the exam, but very few show the score. Here all 40 Paper I questions are set against our 120.
The post has three parts:
- The method: what we wrote before the exam, and the four levels we used to score each question (sections 1 and 2).
- The result: the score with every hit, the misses and extras, and the lesson weights we got right and wrong (sections 3 to 5).
- The lesson: what a model paper can and cannot do, and how to use one (section 6).
1. What we wrote before the exam
We wrote three Part I papers with 40 MCQs each, three days before the exam. Each paper took a different angle:
- Model Paper 1 follows the recent trend. Its lesson mix is closest to the 2025 paper.
- Model Paper 2 follows the plain average of the three new-syllabus papers (2019-22, 2023-24 and 2025).
- Model Paper 3 keeps the recent trend, but fills its slots with syllabus points that no past MCQ had asked.
We wrote three papers because the past papers do not agree on lesson weights. The computer and its potential rose from 2 MCQs to 7, then to 11, across the three new-syllabus papers. Application software fell from 15 to 10, then to 5. One paper cannot cover that whole range, so three papers cover it from three sides.
The three papers cover only Part I, the MCQs. We wrote no Part II questions. So this post scores Part I only.
2. How we scored each question
For each official question, we found the closest question among all 120 of ours. Then we gave it one of four levels:
- Exact: the same point, and the same thing asked. A student who did our question can answer the official one directly. The numbers may differ.
- Close: the same sub-topic and the same idea, but a different angle, device, function or value.
- Topic only: the same lesson area, but a different point inside it.
- Missed: nothing in our three papers touches it.
One more point matters. We set 40 questions against 120, not against 40. A wider net catches more fish. So "Topic only" is weak evidence, because almost any careful set of 120 GIT questions touches most lesson areas. The levels that really count are Exact and Close.
3. The score in numbers
27 of the 40 questions of the GIT 2026 Paper I (G.C.E. (A/L) General Information Technology 2026, Sinhala medium) were Exact or Close. We give each official question in our own English words. Where we give an answer, it is our own worked answer, not official, because the official marking scheme is not out yet.
- Exact: 9 questions (22.5%)
- Close: 18 questions (45%)
- Topic only: 9 questions (22.5%)
- Missed: 4 questions (10%)
So the fair headline is this: 27 of the 40 questions (67.5%) were Exact or Close. The strong claim is smaller: 9 of 40 (22.5%) were Exact. The weak claim is 36 of 40 (90%) by topic. We do not lean on that number, for the reason in section 2.
The 9 Exact hits were spread evenly. Each model paper had 3 of them.
The nine exact hits
- GIT 2026 P-I Q1: the best field for the primary key of a student table. Model Paper 2 Q24 asked the same for a students table.
- GIT 2026 P-I Q2: the correct HTML comment. Model Paper 2 Q38 asked which HTML line is a comment.
- GIT 2026 P-I Q6: the port to connect a printer. Model Paper 3 Q7 asked the port for a flash drive, with the same USB, HDMI and VGA options.
- GIT 2026 P-I Q14: the number of cells in B2:D5. Model Paper 3 Q21 asked the same for B2:D4.
- GIT 2026 P-I Q16: the car park bits 10110 as a decimal number. Model Paper 1 Q9 asked for the decimal value of the same number, 10110. The answer is 22 in both, and it is option (3) in both.
- GIT 2026 P-I Q17: the main job of the ALU. Model Paper 1 Q3 asked which part adds and compares marks.
- GIT 2026 P-I Q19: where the data and instructions now being processed are held. Model Paper 3 Q5 asked the same with the block diagram. The official question leaves some room for doubt, because it also has a "registers" option.
- GIT 2026 P-I Q33: the job of the NIC. Model Paper 2 Q32 asked the same.
- GIT 2026 P-I Q40: ASCII and Unicode. Model Paper 1 Q14 asked why Sinhala and Tamil text needs Unicode.
The close calls
These share the idea, but the official question came from another angle:
- GIT 2026 P-I Q4 secondary storage and GIT 2026 P-I Q13 optical storage: our storage type questions (Model Paper 1 Q2, Model Paper 2 Q4).
- GIT 2026 P-I Q5 the device for a hard copy: our output device question (Model Paper 1 Q1).
- GIT 2026 P-I Q7 system software and GIT 2026 P-I Q9 antivirus as utility software: our disk clean-up questions (Model Paper 1 Q4, Model Paper 2 Q5).
- GIT 2026 P-I Q15 a laptop specification: our computer advert question (Model Paper 1 Q8).
- GIT 2026 P-I Q20 decimal 45 to binary: our decimal to binary questions (Model Paper 1 Q10, Model Paper 2 Q8).
- GIT 2026 P-I Q23 guided and unguided media: Model Paper 1 Q31.
- GIT 2026 P-I Q25 what "docx" is: our file extension question (Model Paper 1 Q17).
- GIT 2026 P-I Q26 why we build a PAN: our Bluetooth earbuds question (Model Paper 1 Q30).
- GIT 2026 P-I Q27 the
$sign for an absolute reference: Model Paper 2 Q19, where the student chooses=A3*$B$1. - GIT 2026 P-I Q28 the formula
=5+3*2^2-8/4: operator order is in all three papers (Model Paper 1 Q20, Model Paper 2 Q18, Model Paper 3 Q20). But none of ours used the^power sign. - GIT 2026 P-I Q29 IoT statements: our IoT example question (Model Paper 1 Q35).
- GIT 2026 P-I Q32 GSM and GPRS: Model Paper 3 Q30.
- GIT 2026 P-I Q36 cropping and sharpening: our crop question (Model Paper 3 Q26).
- GIT 2026 P-I Q37 SVG for vector graphics: our vector logo question (Model Paper 2 Q27).
- GIT 2026 P-I Q38 JPEG, MP3 and PNG statements: our PNG and JPG questions (Model Paper 1 Q27, Model Paper 3 Q27).
- GIT 2026 P-I Q39 raster, vector and resolution statements: Model Paper 1 Q26 and Model Paper 2 Q27.
4. What we missed
Four questions had nothing in our papers:
- GIT 2026 P-I Q3: name the network topology from a picture.
- GIT 2026 P-I Q11: the main job of a wireless base station.
- GIT 2026 P-I Q18: what a multi-core processor allows.
- GIT 2026 P-I Q22: what MICR is used for.
All four misses are hardware and network detail. We gave those slots to other points.
Points we only touched by topic
Nine more questions sat in a lesson area that we covered, but they asked a point that we did not:
- presentation keys: F5 to start a slide show (GIT 2026 P-I Q10) and Ctrl+M for a new slide (GIT 2026 P-I Q12)
- word processing tools: double line spacing (GIT 2026 P-I Q30) and Format Painter (GIT 2026 P-I Q31)
- binary to hexadecimal with 10111100 (GIT 2026 P-I Q21)
- the
altattribute of an image (GIT 2026 P-I Q8) - the main aim of a network (GIT 2026 P-I Q24) and the basic need to set one up (GIT 2026 P-I Q34)
- what the base (radix) of a number system shows (GIT 2026 P-I Q35)
Slots the paper did not use
We also wrote questions on points that the 2026 paper did not ask:
- logic gates and truth tables (one or two in each of our papers)
- ICT in life and society, such as telemedicine, e-government, firewalls and e-waste (two in each paper)
- email Cc and Bcc, and the parts of a URL
- database query and report objects
5. Lesson weights we got right and wrong
We got the weight of three lessons right, wrote too few on two and too many on three. Here is the range across our three papers, against the 2026 paper, by our count:
- The computer and its potential: we wrote 7 to 8, and the paper had 11. Too low.
- The Internet and email (networks and IoT included): we wrote 6, and the paper had 9. Too low.
- Application software: we wrote 8 to 10, and the paper had 8. Right.
- Data representation: we wrote 5 to 6, and the paper had 5. Right in number, but the paper had no logic gate question.
- Multimedia: we wrote 4, and the paper had 4. Right.
- Websites: we wrote 3 to 4, and the paper had 2. Too high.
- Operating systems: we wrote 2 to 3, and the paper had 1. Too high.
- ICT in life and society: we wrote 2, and the paper had none. Too high.
We followed the right trend, but we did not follow it far enough. The computer and its potential went from 2 to 7 to 11, and it stayed at 11 in 2026. Networks rose too. The question format also moved. The 2026 paper had three A, B, C statement questions (GIT 2026 P-I Q29, GIT 2026 P-I Q38 and GIT 2026 P-I Q39). Our three papers had only three such questions in all 120.
6. What this means for students
A model paper predicts the exam only partly. About two in three of the 2026 questions were Exact or Close to a question in our papers. But only about one in four was Exact. And one in ten was not there at all.
The bigger lesson is in how the hits worked. Even an Exact hit gives the method, not the answer. Look at GIT 2026 P-I Q14. Our Model Paper 3 Q21 asked for the cells in B2:D4. That is 3 columns times 3 rows, so 9 cells. The 2026 paper asked for B2:D5. That is 3 columns times 4 rows, so 12 cells. The numbers changed, but the method did not. A student who memorised "9" lost the mark. A student who learned "columns times rows" got it.
Use model papers in this order:
- Sit the model paper as a timed test first.
- For each mistake, find the idea behind the question, not only the answer.
- Read the syllabus point around that idea, so a new angle does not come as a surprise.
- Revise the points that no model paper covered, such as specialist input devices and network topologies.
Try the papers
Sit the real GIT 2026 past paper as a timed test first. Then try GIT Model Paper 1, GIT Model Paper 2 and GIT Model Paper 3, and look for the hits. For the reason behind each 2026 answer, read GIT 2026 MCQ answers explained one by one. After that, the GIT 2025 past paper is the next one to sit.
Past PaperGIT 2026 Past PaperMCQ 40 · මිනි. 60 · ලියූ වාර 16Open paper Model PaperGIT Model Paper 1MCQ 40 · මිනි. 60 · ලියූ වාර 2Open paper Model PaperGIT Model Paper 2MCQ 40 · මිනි. 60 · ලියූ වාර 0Open paper Model PaperGIT Model Paper 3MCQ 40 · මිනි. 60 · ලියූ වාර 0Open paper