When 70 Lakh Students Block Justice for 1.27 Lakh: Delhi HC's CBSE OSM Ruling

A Court Faces an Impossible Arithmetic

On June 12, 2026, the vacation bench of the Delhi High Court heard an urgent plea from the National Students' Union of India. The petition, filed as a public interest litigation, asked the court to direct CBSE to reopen its online revaluation portal for Class 12 answer books evaluated under the On-Screen Marking (OSM) system.

The request seemed reasonable. By the board's own portal data, 1.27 lakh students had filed formal complaints about their evaluation outcomes. The grievances were specific and documented: blurred scans, missing pages, mismatched answer sheets, unmarked questions, and marks that students claimed were implausibly lower than their preparation warranted. CBSE had already reduced the revaluation fee to ₹100 — an implicit acknowledgment that the standard ₹500 barrier was inappropriate given the scale of complaints — and relaunched the portal on June 2. That window had since closed.

The court declined to grant the direction.

The bench did not dismiss the students' concerns. It cited a more fundamental constraint: reopening the portal, even for a week, would affect more than 70 lakh students who appeared in the Class 12 examination — students whose admission processes to undergraduate courses were already underway. The Solicitor General, appearing for the government, made the arithmetic plain. For every day the revaluation portal stays open, the declaration of final corrected marks gets delayed. For every day that final marks are delayed, thousands of university admission offers hang in limbo.

The matter was listed before the roster bench for July.

The Cascade Nobody Planned For

This is the part of large-scale digital evaluation rollouts that procurement committees and technology vendors rarely model: the downstream cascade when something goes wrong.

CBSE's Class 12 evaluation involved approximately 16 million answer scripts scanned and uploaded to the Onmark platform. The OSM system assigns individualised question-level bundles to evaluators, who mark on screen rather than on physical paper. The intention was to eliminate transit damage, ensure double-valuation, and create granular digital records.

What the rollout did not adequately account for was the scanning quality layer. Physical answer sheets — written by students in examination centres across 36 states and union territories — must be digitised before they enter the digital evaluation pipeline. Scanning quality depends on the equipment at each centre, operator training, sheet handling, and environmental conditions. A single scan failure translates directly into a digital evaluation failure that no back-end technology can retrospectively correct.

The scale is relevant. If even 1% of 16 million scanned sheets had scanning anomalies, that is 1.6 lakh problem scripts. The 1.27 lakh complaints on the CBSE portal indicate the actual incidence was in this range.

Once those scripts entered the evaluation system with defects — blurred or missing pages — evaluators encountered incomplete data. Some gave zero marks for questions they could not see. Some skipped pages. The digital record preserved this flawed evaluation exactly: timestamped, evaluator-tagged, and audit-trailed, but factually wrong.

When students later accessed their question-wise marks, the anomalies became visible in a way they never had under paper evaluation. OSM, paradoxically, made errors more legible even as it made them harder to correct at scale.

What the Court's Logic Reveals

The Delhi HC's June 12 reasoning encodes a structural truth that any institution planning digital evaluation at scale needs to absorb: at sufficient scale, the cost of remediation can exceed the system's capacity to administer it without harming third parties.

Under paper-based evaluation, a student who suspected an error could request a physical revaluation within a narrow window. The process was slow and expensive, but it was bounded — each revaluation request touched only that student's script, and downstream effects on other students were minimal.

Under OSM, when 1.27 lakh students have legitimate grievances, the revaluation process is no longer bounded. Reopening the portal triggers a cascade: CBSE must process 3.86 lakh answer copies (the figure cited in court), each requiring a second evaluator session, quality review, and mark reconciliation. The results must flow back into the aggregate marksheet system before any downstream institution — a central university, a state admission authority, a scholarship board — can finalise its merit list.

The 70 lakh students who had no complaints suddenly find their admission timelines held hostage to the revaluation workload generated by 1.27 lakh others. This is not a trivial trade-off. For students applying to undergraduate programmes that fill seats in a single counselling round, a four-week delay in final marks can mean losing a programme entirely.

The court chose the lesser harm. The real question is: why was this choice necessary?

What Zero-Downtime Revaluation Architecture Looks Like

The CBSE OSM controversy has generated extensive commentary on what went wrong. Less attention has been paid to what a correctly designed system would look like — one that could absorb 1.27 lakh revaluation requests without cascading into a system-wide suspension.

Three design principles distinguish systems that can absorb this load from those that cannot.

Scanning validation at ingest, not at dispute. A system that catches scanning anomalies before scripts enter the evaluation queue prevents the downstream problem. Every uploaded image should pass an automated quality gate: minimum resolution, page count validation, student identity match, presence of QR or barcode markers. Scripts that fail the gate should be flagged for rescanning before evaluation begins, not after a student files a complaint three months later.

Parallel revaluation as a default, not an exception. OSM systems can assign a fresh evaluator to a disputed script within hours, running the new evaluation in parallel with the original. The mark difference can be computed in real time, and moderation triggered automatically if the difference exceeds a threshold. This does not require reopening a portal to all 70 lakh students — it requires a separate moderation queue for flagged scripts that runs continuously through the result cycle.

Decoupled timelines. Final mark declarations can be structured to release a primary certificate while flagged scripts remain in review. A student without a complaint receives their final marks on time. A student with an active complaint receives a provisional score with a committed resolution window. This decoupling prevents the 70 lakh versus 1.27 lakh trade-off from ever reaching a court.

The Three Numbers Every Institution Should Know

Before any institution implements on-screen marking at scale, three numbers should be established in writing with the technology vendor.

Scanning rejection rate. What percentage of uploaded scripts are expected to fail the automated quality gate, and what is the process for rescanning? A vendor that cannot answer this question has not modelled the scanning failure scenario.

Revaluation throughput. How many scripts can the platform process through a second-evaluation cycle per day? If the answer is lower than 1% of the total script volume, the system cannot absorb a realistic complaint rate without timeline disruption.

Result decoupling support. Does the platform support releasing results for the uncontested population while contested scripts remain in the moderation queue? If the answer is no, the institution is structurally locked into the same 70 lakh versus 1.27 lakh dilemma CBSE faced.

These are not advanced technical requirements. They are foundational operational requirements that define whether a digital evaluation system is designed for resilience or for best-case scenarios only.

What the July Hearing May Decide

The Delhi HC roster bench, hearing the matter in July, will face an additional complication: by that point, the most competitive undergraduate admission cycles will have completed their first rounds. Correcting marks for students who missed cutoffs by one or two marks due to OSM errors may no longer benefit them materially in the 2026 admission cycle.

This is the downstream cost of evaluation failures that are not detected and corrected at source: the remedies available after the fact are necessarily inferior to the harm suffered. A student whose marks were understated in May and whose revaluation result arrives in August has, in functional terms, lost a year.

The court cannot undo that loss. What it can do — and what the July hearing may address — is establish a framework for what CBSE must do differently before the 2027 examination cycle begins.

For institutions implementing their own digital evaluation systems, the lesson is available now, before the 2027 cycle, before a July hearing is necessary.

Related Reading