★★★★★
"The one step solution for my son. Modern Age Coders make learning coding so simple that kids love it. The teachers explain complex concepts clearly with practical exercises and interactive content."
Ria Mukherjee
Parent
CBSE Class 10 · Artificial Intelligence, code 417 · Marking scheme explained, 2026-27
The class 10 AI marking scheme that CBSE gives examiners lists value points: the ideas an answer must contain, and how many marks each is worth. We read the official schemes for the 2025 and 2026 board papers and the 2025-26 sample paper, and describe below, in our own words, how they split the marks. Then each kind of question is answered in full, annotated half mark by half mark, beside a weaker version that loses marks. In our live batch, every mock is marked this way on screen.
Live online · marking rules in plain words · annotated model answers · weak answers marked down · mocks marked to the scheme · first class free
Start here
The 417 course whose batch writes and marks these answers, the Python course for the practical half the theory scheme does not cover, and the Orange course for the no-code tools named in objective items.

The batch / 417 Class 10
The full 417 course, with written answers marked against the value points after every unit and in each full mock.
Open the course →
Practical half / Python
The 50 practical marks have their own assessment: NumPy, Pandas and Matplotlib programs for the file and the practical exam.
Open the course →
No-code tools / Orange
The no-code workflows behind several one-mark items and the practical units, built and explained.
Open the course →Ans. The short version
The class 10 AI marking scheme for CBSE 417 gives examiners value points, not full answers: any correct wording earns the mark if the point is there. In the 2025 and 2026 schemes, objective items carry 1 mark each with no negative marking, 2-mark answers usually split into two 1-mark points or a point plus two half-mark examples, and 4-mark answers into named parts such as 1 per step or half a mark per formula and per substitution. Modern Age Coders marks practice answers this way live in a Monday and Wednesday 9:00 PM IST batch, or one to one.
Q1. How is the class 10 AI paper marked?
The 417 theory paper is 50 marks in two hours: 24 objective marks in Section A and 26 written marks in Section B. The 2025 and 2026 board schemes and the 2025-26 sample paper scheme all use this shape.
| Part of the paper | Choice | Marks each | Total | How the scheme marks it |
|---|---|---|---|---|
| Section A, employability skills | any 4 of 6 | 1 | 4 | Right or wrong; often an option letter or its text is accepted |
| Section A, AI items (four questions) | any 5 of 6 in each | 1 | 20 | Same; no negative marking |
| Section B, employability skills | any 3 of 5, 20 to 30 words | 2 | 6 | Usually 1 per relevant point |
| Section B, AI short answers | any 4 of 6, 20 to 30 words | 2 | 8 | Two points, or a point plus examples |
| Section B, AI long answers | any 3 of 5, 50 to 80 words | 4 | 12 | Split into named parts, steps or calculations |
Examiners look for the points, so make each point easy to find: one idea per line, the key term written exactly (precision, data acquisition, justice), and the example labelled "Example:". An answer that is right but buried in a paragraph risks a point being missed.
A note on sources: we read set 4 of the 2025 and 2026 papers and the sample paper scheme. Other sets have other questions but the same marking style. The 2025 paper still carried some older topics, such as AI bias and data science applications; 2026 follows the current units. Unit-by-unit weightage is on the exam date and paper pattern page.
Q2. How are one-mark items marked?
Every objective item is one mark, all or nothing, with no negative marking. The schemes accept either the option letter or the option text, and some items list more than one acceptable answer.
The 2026 scheme writes answers as "(C) OR the option's text". Writing both is safest: a letter alone can be misread, and the words show you knew it.
In 2026, a two-statement item about overfitting accepted two options, because each statement was true on its own. In 2025, a chatbot item accepted several names for a smart bot.
The 2025 chatbot item gave half a mark when a student named only an example product instead of the type. So even a one-mark item can be partly credited when the scheme allows it.
Thirty-six more objective items with answers are on Class 10 AI MCQ questions.
Q3. How are 2-mark answers split?
A 2-mark answer is almost always worth two separate points. The schemes use four splits again and again, and knowing which one a question uses tells you what to write.
| Split | Where we saw it | What to write |
|---|---|---|
| 1 + 1, two points | Two characteristics of a classification model (2026); two stages explained (2025-26 SQP) | Two separate lines, each a complete point |
| 1 for a difference + ½ + ½ for examples | Sector-based against value-based frameworks (2026) | One clear difference, then one example of each, labelled |
| 1 for a difference + ½ + ½ for applying it | Stemming against lemmatisation on a plural word (2026) | The difference, then the stem and the lemma of the given word |
| Graded by how many points | The 7 Cs of communication (2026): all seven for 2, fewer for 1½, 1 or ½ | Every item in the list, even if short of time |
Full-mark answer. Classification predicts a label or category from fixed classes, while regression predicts a continuous number [1]. Example of classification: sorting emails into spam or not spam [½]. Example of regression: predicting tomorrow's temperature in degrees [½].
Answer that gets 1. "Classification puts things in groups and regression is for numbers, e.g. house price." The difference earns its mark; there is only one example, and it is not labelled as regression, so at most half more is likely.
Full-mark answer. Model evaluation is checking how well a trained model performs, using metrics such as accuracy, precision and recall on data it has not seen [1]. It is needed because the scores show the model's weak points, so it can be improved before people rely on it [1].
Answer that gets 1. "Evaluation is the fifth stage of the AI project cycle where we test the model." That places it, but gives no reason it is needed, so the second point is missing. This "define, then explain" split is the one the 2025-26 sample paper used for its evaluation item.
Every two-mark difference pair the units offer is answered on CBSE Class 10 AI difference between questions.
Q4. Is there step marking in Class 10 AI?
Yes, in effect. The four-mark items are split into parts, steps or calculations, each with its own mark, so a student who gets the final value wrong can still earn most of the marks for the matrix, the formulas and the substitution.
Q. A library tests a model that predicts whether a borrowed book will be returned late, on 400 loans. It predicted "late" for 80 loans, of which 72 were late. Of the 320 it predicted "on time", 28 were late. (a) Draw the confusion matrix. (2) (b) Calculate precision and recall. (2)
(a) Reality: Late Reality: On time
Pred: Late TP = 72 FP = 8
Pred: On time FN = 28 TN = 292 [2]
(b) Precision = TP / (TP + FP) [½]
= 72 / (72 + 8) = 72 / 80 [½]
= 0.90
Recall = TP / (TP + FN) [½]
= 72 / (72 + 28) = 72 / 100 [½]
= 0.72
Writing each formula on its own line, then the substitution, gives the examiner four separate places to award half marks.
| Slip | Likely cost |
|---|---|
| Matrix drawn with all four cells right | 2 of 2 |
| Matrix with only two cells right | about 1½ |
| Matrix frame with no values | ½ |
| FP and FN swapped, so "precision" = 0.72 | loses cell marks; formula marks can survive, and the error is penalised once |
| Formula only, no substitution | ½ per part |
| Right formula, arithmetic slip in the last line | usually little or nothing |
The costs above follow the 2026 and 2025 schemes for their matrix items: two marks for a fully correct matrix with partial credit below that, and half marks for each formula and its substitution. Both schemes add a note that the calculations themselves can be ignored, which puts the weight on the method. Still write the final value: accuracy here is 364 / 400 = 0.91, and F1 is 2 x 0.90 x 0.72 / 1.62 = 0.80, not the plain average 0.81.
Q. Apply the four steps of Bag of Words to: D1 "Robots help farmers." D2 "Farmers teach robots and robots help." (4)
Step 1, normalise [1]: D1 = robots, help, farmers; D2 = farmers, teach, robots, and, robots, help.
Step 2, dictionary [1]: robots, help, farmers, teach, and.
Step 3, vector for D1 [1]: [1, 1, 1, 0, 0].
Step 4, table for all [1]: D1 [1, 1, 1, 0, 0]; D2 [2, 1, 1, 1, 1].
The 2026 scheme gave one mark per step and said steps 3 and 4 may be shown together. A table without the step names risks the marks for the steps you skipped. Fully worked versions are on Bag of Words and TFIDF for Class 10 AI.
The bioethics case answer is worked mark by mark on AI project cycle and ethics.
Q5. Where are marks usually lost?
Most lost marks in this paper are not lost on hard topics. They go on answers that miss one value point the scheme was looking for. These are the patterns we see most when marking mocks.
| What happens | Which split it breaks | The habit that fixes it |
|---|---|---|
| A difference written, but no examples | Difference plus two half-mark examples | Write "Example:" twice whenever a question says "with an example" |
| Principle described but not named | Half for the name, half for the explanation | Start each line with the principle's name: "Justice: ..." |
| Only the final number in a numerical | Half per formula and per substitution | Formula line, substitution line, answer line, every time |
| Bag of Words table with no steps | One mark per named step | Head each part "Step 1", "Step 2" and so on |
| Two points merged into one sentence | 1 + 1 | One point per line; never let a "because" hide the second point |
| Long answers past the word limit | Not a rule for marks, but time lost | 20 to 30 words for 2 marks, 50 to 80 for 4; stop when the points are made |
| Extra questions attempted "just in case" | Only the higher answer counts | Choose early, spend the time checking |
The marking scheme covers the 50 theory marks only. The other 50, the practical file, practical exam, viva and project, are assessed in school; that half is on the project and practical file page. Past papers to practise these splits on are described on CBSE Class 10 AI previous year question papers.
Q6. How do classes use the marking scheme?
Knowing the splits is half the work; the other half is writing to them under time until it is automatic. That only happens when someone marks your answers the way the board will and shows you the half marks you dropped.
On Mondays students write two-mark and four-mark answers against the clock, to the 20 to 30 and 50 to 80 word limits the paper sets.
On Wednesdays the teacher marks a few answers on the shared screen, writing the half marks in the margin, so everyone sees which line earned what and which point was missing.
Each student rewrites one answer that lost marks. From December, full timed mocks are marked against the value points and returned with a mark per question.
The mock calendar is on the sample paper and mock tests page. Students can pause any marking session to argue for a half mark; working out why the scheme would or would not give it is often where the lesson lands.
Batch or personalised: you choose
In the batch, answers from the whole class are marked on screen, so students learn from each other's lost half marks as well as their own. One to one marks only your child's answers and rewrites them until every value point is there. Same teachers, same scheme; monthly billing, no admission fee, stop at the end of any month.
Option 1 · Fixed batch timings
₹1,499
per month, two live classes a week
Option 2 · Choose your own timing
₹4,999
per month, one private class a weektwo private classes a week, at a time set with you
A student who knows the content but scores below it in school tests often has a marking-scheme problem, and a few one to one sessions on answer layout tend to fix it. A student still learning the units gains more from the batch, where the content and the answer writing are taught together. Fees are shown in your own currency.
Studying outside India?
Answer scripts from CBSE schools abroad are marked to the same scheme as those from India. Here is when the evening batch falls in your time zone this winter.
| Where you live | Mon and Wed 9:00 PM IST | Best format |
|---|---|---|
| India (IST) every Indian city | 9:00 PM | Batch fits |
| UAE and Oman Dubai, Abu Dhabi, Sharjah, Muscat | 7:30 PM | Batch fits |
| Saudi Arabia, Qatar, Kuwait, Bahrain Riyadh, Jeddah, Dammam, Doha, Kuwait City, Manama | 6:30 PM | Batch fits |
| Nepal Kathmandu | 9:15 PM | Batch fits |
| Ethiopia and East Africa Addis Ababa and East Africa Time | 6:30 PM | Batch fits |
| Singapore and Malaysia Singapore, Kuala Lumpur | 11:30 PM | Choose 1-on-1 |
If 9:00 PM IST is too late where you live, one to one runs at an hour you choose. More for families outside India is on CBSE Class 10 AI online classes in the Gulf and abroad.
What families say
Real reviews from real families. We neither write nor commission them.
★★★★★
"The one step solution for my son. Modern Age Coders make learning coding so simple that kids love it. The teachers explain complex concepts clearly with practical exercises and interactive content."
Ria Mukherjee
Parent
★★★★★
"Modern Age Coders has been a game-changer for me. I struggled to grasp IT concepts and coding before joining, but their classes transformed everything. I can now confidently write complex programs with ease."
Samriddha Mondal
Student
★★★★★
"One of the most wonderful education centres out there. Education is not limited to school syllabus but focuses on skill development."
Vansh Agarwal
Student
★★★★★
"My child Dhairya is really enjoying the Modern Age Coders classes. This is his first online class and he eagerly looks forward to it. I can already see his improvement, and the teachers are very cooperative."
Sonam Oswal
Parent of Dhairya
★★★★★
"Modern Age Coders have wonderful teachers who teach in a clear, easy and practical way. The teacher boosts students' confidence and inspires them to learn without hesitation."
Sonu Goyal
Parent
★★★★★
"I highly recommend this computer coding class! The teachers are incredibly knowledgeable and passionate about coding."
Ritu Kedia
Parent
The rest of the Class 10 series
Every page in the CBSE Class 10 AI (417) series, grouped by what you need: classes, each unit, question practice, notes and revision, and the exam itself.
ICSE Class 10 Computer Applications (Java): The Tue and Wed ICSE Java batch ICSE Computer Applications syllabus ICSE Java programs practice ICSE CA important questions
Questions about the AI marking scheme
It is the document CBSE gives examiners for each paper, listing the value points an answer should contain and how many marks each point earns. For 417 it covers the 50-mark theory paper. CBSE publishes main exam schemes on its website and a scheme with each sample paper, which is how students can see how answers are marked.
CBSE posts main exam marking schemes on cbse.gov.in, by year and subject code, and sample paper schemes on cbseacademic.nic.in. Each main exam has several sets with different questions; we read set 4 of the 2026 paper, so other sets differ in questions but not in marking style.
In effect, yes. Four-mark numericals and Bag of Words items are split into parts: marks for the confusion matrix, half marks for each formula and each substitution, or one mark per Bag of Words step. The 2025 and 2026 schemes also note that calculations can be ignored, so the method carries most of the marks.
No. The 2025 and 2026 papers and the 2025-26 sample paper all state there is no negative marking in Section A. Attempt every objective item you choose, and use elimination when unsure.
Usually as two separate points of one mark each, or one mark for a difference plus half a mark for each of two examples. Some list-based answers are graded by how many items you give. Write each point on its own line so the examiner can find it.
No. The scheme says its value points are guidelines and that a student's own expression earns the marks if it is correct. Key terms still matter: write precision, data acquisition or justice exactly, because those words are what the examiner looks for.
The examiner marks the extra answer too, but only the higher-scoring one is counted and the other is struck out. It rarely helps, because the time is better spent checking the answers you chose.
The 417 paper asks for 50 to 80 words for four-mark answers and 20 to 30 words for two-mark answers. The marks come from the value points, not the length, so aim to cover every point clearly within the limit.
No one can say until CBSE publishes the 2026-27 sample paper for 417, and the one listed so far is the 2025-26 paper. The 2025-26 sample paper and the 2026 board paper used the same structure and splits, so practising to them is the sensible preparation.
Yes. Answers are written to time and marked against the value points in ordinary classes, in the batch or one to one, covered by the monthly fee shown on this page in your own currency. There is no admission fee and the first class is a free demo.
Start
In the free first class your child writes one two-mark and one four-mark answer, and the teacher marks both live, half mark by half mark, against the way the board splits them. You will see where marks go missing, and the mentor will tell you honestly what to work on and whether the batch or one to one fits better.
Want the full picture of the subject first? Read the 417 syllabus explained.
WhatsApp us · +91 91233 66161 · connect@modernagecoders.com
Every class is live video. The form requests a callback and reserves nothing.
A full live class of about 45 to 60 minutes at a slot you pick, with a written skill report and learning roadmap on WhatsApp afterwards. The fee is adjusted against your first month if you enrol.
Book a priority demo