How BAND9AI Evaluates IELTS Reading
Reading · AI scoring · May 2026
Direct answer
BAND9AI evaluates Reading the same way IELTS does on paper: each of the 40 answers is marked right or wrong against the key, raw score converts to a band, and misses are clustered by question type. AI does not invent a "comprehension band." It surfaces whether you lost marks on TFNG, matching headings, sentence completion, or time bleed on Passage 3—so practice targets the leak, not generic "read more."
Three-step Reading evaluation
Mark Key match per item—spelling and word limits enforced
Convert Raw correct → official Reading band chart
Diagnose Misses tagged by item type and passage
Question-type clusters that matter
| Cluster | Typical leak |
|---|---|
| TFNG / YNNG | NOT GIVEN vs FALSE confusion |
| Matching headings | Main idea vs detail in paragraph |
| Summary / gap fill | Grammar fit + word limit |
| Passage 3 timing | Unfinished items, not ability |
Limits vs human tutoring
AI cannot replace skimming strategy coaching in the room. Use Reading weakness tools for scored mocks, then drill one cluster per week. Cross-check with chronological order traps if history/science passages stall you.
Key takeaways
- Reading band = correct answers on the 40-item chart.
- Tag misses by question type, not only total wrong.
- TFNG and matching headings deserve separate drill blocks.
- Passage 3 unfinished items often mean timing, not vocabulary.
FAQ
No. Reading is answer-key based. Band comes from correct count on the 40-item chart; AI adds question-type and paraphrase-gap diagnosis.
Same scoring logic—different passages and item mixes. GT Section 3 still maps to the same raw-to-band conversion.
Yes. Misses on TFNG, matching headings, and summary completion are tagged separately so you drill the right skill.
See which Reading question types cost you marks—not just your raw score.
Get IELTS Reality Check →