IELTS Writing Benchmark 2026

IELTS Writing Checker recorded 3,222 completed IELTS evaluations by 21 August 2026. We filtered those records for the correct site and exam, plausible scores from 1 to 9, a minimum response length, and exact duplicate text. The resulting research corpus contains 1,941 responses from 1,481 learner accounts.

Academic Task 2 makes up more than half of the cleaned corpus. Academic Task 1 contributes a further 709 responses. The General Training Task 1 and Task 2 cohorts contain 76 responses each, so we report their sizes but do not draw detailed conclusions from them.

In both large Academic cohorts, Grammatical Range and Accuracy had the lowest mean stored criterion estimate. That is a finding about this self-selected practice corpus and the checker's historical estimates. It is not evidence that grammar is the weakest criterion among all IELTS candidates.

People who choose an online writing checker are a self-selected group. This dataset should not be treated as representative of all IELTS candidates, countries, first languages, or test centres.

A learner can contribute more than one unique response, so the 1,941 observations are not all statistically independent. Exact-text deduplication removes identical responses but does not identify lightly edited versions.

Historical evaluation rows do not consistently record the evaluator model or prompt version. Scores and criteria in this report are therefore described as stored practice estimates, not official bands. No raw essays were reviewed or published for this report.

The General Training cohorts are small. Their counts are included for completeness, but the criterion finding is limited to the two larger Academic cohorts.