Intelligence / The benchmarks & the desk / Bench Intelligence
Every fund prices bench delay informally and nobody publishes it. This is the median time from admission to verdict at each NCLT bench, split by the verdict it ends in, by who filed and by the size of the claim — with the caseload still waiting, the money the wait costs, and the frictions that make it long.
separate the slowest bench from the fastest. Allahabad takes a median 663 days to a verdict; Guwahati takes 350. Same code, same 330-day clock.
Days against the national median in the tiles above — red is slower, green faster; the bar is the bench's own median, which the register below prints. A bench with fewer than 30 dated verdicts shows its n instead — below that a median is noise wearing a number. The 10 benches beyond the five busiest are on the member desk.
is what an approved resolution plan takes over a liquidation order, nationally — 631 days against 424. Rescue is the slow exit everywhere, and it holds at 8 of the 9 benches where both sides clear the guard.
The exception on the record is Allahabad, where the liquidation order takes longer than the plan. Each bench's own two clocks are inside its row below.
is the median age of a CIRP still running, counted from admission. 2,054 of them are already past the 330-day statutory limit and 1,165 past three years. The caseload is in the tiles above.
Age since admission on every live CIRP the bench is carrying, n=2,666. This is the only figure on the page that is not a verdict: it is the wait that has not ended yet, and it is the one a creditor is actually sitting in.
The bench desk the five busiest benches public · the full comparison for members
Narrow by bench, by the verdict the case ended in or by the year of that verdict; the benchmark line recomputes for the selection against the national record. Caseload and the verdict split are whole-record columns; the median and the 330-day share follow the filter.
The median CIRP took 202 days to a verdict in 2017 and 632 days across the last three years — a national figure, on n=1,619. A verdict passed in an early year could only be on a case admitted months before it, when the code itself was a year old, so the first bars are structurally short; the last five years are the comparison that carries weight.
Median days from admission to verdict, by the year the verdict was passed. Red is beyond the national median in the tiles above, green at or inside it.
The 330-day limit is the statute’s own answer to the question this page asks; the money is what the answer is worth. Both are national figures here, and the bench-level record is in the register above.
A creditor who waits the national median gets back a median 22.9 paise in the rupee where a plan is approved, and a median 2.0 paise where the company is liquidated and the estate finally dissolved. That is the real spread on this page: not the months between the fastest bench and the slowest, but the verdict the months end in.
The frictions are the part of the record nobody publishes: how often a committee changed its professional, how often the clock had to be extended, how often an admission was set aside on appeal and how many times a company went to market before anyone bid. Counts over the whole order feed to 18 Sep 2026.
A second Form G round is the clearest single sign that a case will end in liquidation rather than a plan: the committee went back to the market and still found no offer it could vote on. 33 plans were refused by a bench after the committee had approved them.
Members read each bench’s own trend against its own history, the haircut its plans carry and the paise its dissolved estates returned, and the frictions — Form G rounds, RP replacements, extensions, admissions set aside — ranked across all 15 benches.
Bench speed is one input, and the honest use of it is triage: where to look first, not what a case will take. A median here is a historical fact about cases that have already ended — it is not a forecast for any live matter, and the mix of cases a bench hears moves it as much as the bench does.
A bench hearing more small operational-creditor petitions will look faster than one hearing large financial-creditor matters, because a case under Rs 50 Cr closes a median 184 days sooner than one over Rs 1,000 Cr. Each row’s own split by claim band is the control.
The median above is over cases that ENDED. The cases still running past the statutory limit never enters it, which is why the register carries the caseload and its age beside the median rather than under it.
Withdrawals are out of every timing figure on this page: a Section 12A settlement exits early by design and would flatter a slow bench. They stay in the verdict split, where they are an outcome and not a duration.
As of NCLT/NCLAT orders to 18 Sep 2026 · IBBI publications and the Form G register · bench attribution as stated in the source documents · medians are historical, over the record since 2017, and are not a forecast for any live case.
MethodologyGlossaryReport an error corrections are archived with their before-value and shown on the record.