IB World Religions SL IA Grader
AI grading for the IB World Religions SL Internal Assessment: predicted grade band and per-paragraph feedback. The World Religions investigative study...
The World Religions investigative study rewards a clear rationale, a workable plan, and critical reflection on findings from a religion as it is practised, including insider and outsider perspectives.
What we grade for
- Rationale and preliminary research
- Plan for study
- Significant findings
- Critical reflection and evaluation
- Referencing and format
Common pitfalls
- Rationale thin or generic
- Findings described without reflection
- Format or word limit not followed
How the World Religions SL IA is marked
- Rationale and preliminary research (max 8) - Why the topic, and what the early research showed
- Plan for study (max 3) - A workable plan for the investigation
- Summary of significant findings (max 6) - What the study found, in your own words
- Critical reflection and evaluation (max 10) - The findings and the method weighed, with insider and outsider views
- References and compliance with format (max 3) - Sources referenced; the required structure and word limit
Frequently asked questions
What is the MyRevisionAgent Internal Assessment grader?
An AI grader that scores your IB Internal Assessment against the real, subject-specific IA rubric - /14 for Economics, /24 for the sciences, /40 for English A, /25 for History, and so on. You get a predicted mark on that scale, a skills breakdown of where the marks went, feedback mapped to the rubric, and clause-level annotations overlaid on your PDF.
Which IA subjects does it support?
34 subject-levels across HL and SL: Biology, Business Management, Chemistry, Computer Science, Economics, English A Lang & Lit, English A Literature, Environmental Systems and Societies, Geography, German A, Global Politics, History, Mathematics AA, Mathematics AI, Philosophy, Physics, Psychology, and World Religions. Each subject-level is scored on its own IB scale, with feedback mapped to that subject's rubric rather than a generic mark scheme. Subjects with few or no training essays on their current rubric - including Psychology and Computer Science on the May 2027 rubrics - show a wider error range and a note on the result.
Why is the maximum mark different for each subject?
Because the IB awards IAs out of different totals depending on the subject. Economics is /14, sciences are /24, Computer Science is /30 (from May 2027), English A (Language & Literature and Literature, HL and SL) is /40, History is /25, Maths is /20. Our model knows the correct denominator and rubric weights for each subject-level and grades against them.
How accurate is the AI grading compared to a real IB examiner?
Our model is trained on 1,678 real, IB-awarded Internal Assessments and ships every prediction with a confidence range. On the same 376 held-out essays used for the head-to-head comparison, our grader lands within one grade band 90% of the time (92% across the full four-seed cross-validation) versus 54% for ChatGPT (GPT-5.5), and matches the exact grade band 2.6× as often; Claude Opus 4.8 reaches 77% within one band - none were trained on the IB rubric or on IB-awarded marks.
Is the IA grader free?
We add 2 free grading tokens to every new account on signup, usable on any grader including the IA. After that a grade is £9.99 pay-as-you-go (cheaper in packs), or £25.99 a month for five grades on the plan. There is no card required to sign up and explore.
Is AI grading allowed by the IB? Will it count as academic misconduct?
Using AI feedback on your own work is treated the same as feedback from a teacher or tutor - the IB's academic-integrity policy targets AI-written submissions, not AI feedback on your own draft. We never write your IA. We grade it and explain why so you can revise. Always disclose AI tool use as instructed by your school.
Do I get per-paragraph feedback on my IA?
Yes. Every flagged clause is highlighted on your IA in the colour of the skill it concerns - focus & RQ, analysis & argument, evidence & method, knowledge & terms, structure & clarity - with a severity label and a specific revision note. Clauses that already earn their marks get a green check. Click any highlight to jump to its note.
Does the grader handle data tables, graphs and figures?
Yes. Grading reads your PDF's layout directly, so data tables, equations and figure captions are captured and analysed as part of methodology. Heavily image-driven IAs (e.g. Visual Arts) aren't supported yet.
How long does grading my IA take?
Grading takes 4-6 minutes end-to-end. Longer, image-heavy IAs sit toward the upper end of that range, since both the PDF extraction and the feedback pass scale with document length.
How is this different from grading an Extended Essay?
IAs are subject-bound to a specific rubric and shorter than EEs. We use a Ridge regression model with a one-hot feature for each trained subject-level and 50 input signals (subject-aware LLM features plus deterministic signals like citation density and methodology length), so an Economics commentary isn't graded with the same weighting as a Physics investigation.
How is this different from asking ChatGPT or Claude to grade my IA?
A general-purpose LLM has never seen the IA mark scheme or thousands of real IB-awarded IAs. It can write convincingly about your work but its score is a guess. Our grader uses a frontier LLM to read your IA and rate it on rubric features, then a ridge-regression model trained on 1,678 real IB-awarded IAs turns those ratings into the mark - the LLM never picks the mark itself. Every grade carries an explicit uncertainty range.
Is my IA used to train your models?
No. Submissions are never used for training. They are processed to produce your grade and stored only against your own account history so you can revisit them.