AI vs Last-Day GMP - How Accurate Are Our Listing-Gain Calls?
Every listing below compares our AI-predicted listing gain with the last-day Grey Market Premium(the GMP quoted on the IPO close date, which is what most retail investors actually see). We track the absolute error against the real listing-day close. The AI doesn't claim to be perfect - it misses on euphoric hype IPOs too - but on average it lands much closer to reality than last-day GMP.
Average Prediction Accuracy - AI vs Last-Day GMP
Higher = closer to actual listing gainAccuracy = 100% minus the average absolute error vs the actual listing-day gain. Last-day GMP tends to overshoot on euphoric IPOs (Denta Water, Standard Glass, Waaree) and undershoot on quiet mainboards (Hyundai, Ola Electric), which drags its average. Our AI isn't immune - it famously under-called Bajaj Housing and Premier Energies - but it stays much closer to the listing-day truth on average.
Recent Mainboard Listings - Head to Head
Last 12Mainboard IPOs. "AI Err" and "GMP Err" are the absolute gap vs the actual listing-day gain - lower is better. Last-Day GMP is the grey-market premium quoted on the IPO close date.
| IPO | Listed | Actual | AI Pred | AI Err | Last-Day GMP % | GMP Err |
|---|---|---|---|---|---|---|
KI Kissht Mainboard · Digital Lending & Fintech | 2026-08-05 | +11.1% | +4.6% | 6.5% | +2.9% | 8.2% |
OP Om Power Transmission Mainboard · Power Transmission / EPC | 2026-04-17 | +6.3% | +1.2% | 5.1% | +1.1% | 5.2% |
PO Powerica Mainboard · Power Solutions / Energy | 2026-04-02 | -7.3% | -13.9% | 6.6% | +1.0% | 8.4% |
SP Sai Parenterals Mainboard · Pharmaceuticals | 2026-04-02 | +2.0% | +0.4% | 1.6% | +0.0% | 2.0% |
AB Aeroplane Basmati Rice Mainboard · FMCG / Rice Export | 2026-04-02 | -5.7% | -10.2% | 4.5% | +0.9% | 6.6% |
CM Central Mine Planning Mainboard · Mining Consultancy / PSU | 2026-03-30 | -7.0% | -8.9% | 1.9% | +5.2% | 12.2% |
GC GSP Crop Science Mainboard · Agrochemical | 2026-03-24 | +2.5% | -4.1% | 6.6% | +0.0% | 2.5% |
IN Innovision Mainboard · Manpower & Facility Services | 2026-03-23 | -9.9% | -10.9% | 1.0% | -4.8% | 5.1% |
RS Rajputana Stainless Mainboard · Stainless Steel Manufacturing | 2026-03-19 | +0.0% | -5% | 5.0% | +0.8% | 0.8% |
SM SEDEMAC Mechatronics Mainboard · Auto Electronics / ECU | 2026-03-11 | +13.5% | +7.1% | 6.4% | -1.5% | 15.0% |
OE Omnitech Engineering Mainboard · Precision Engineering | 2026-03-05 | -11.0% | -17.4% | 6.4% | +1.8% | 12.8% |
PR PNGS Reva Mainboard · Jewellery Retail | 2026-03-04 | -2.8% | -5.5% | 2.7% | -0.5% | 2.3% |
Year-over-Year Accuracy
Full Prediction Log
All 12 listed IPOs in our dataset, newest first. New listings are auto-ingested from our scraper pipeline.
| IPO Name | List Date | Exchange | AI Pred | Last-Day GMP % | Actual | AI Err | Accuracy | Result |
|---|---|---|---|---|---|---|---|---|
KI Kissht | 2026-08-05 | Mainboard | +4.6% | +2.9% | +11.1% | 6.5% | Miss | |
OP Om Power Transmission | 2026-04-17 | Mainboard | +1.2% | +1.1% | +6.3% | 5.1% | Miss | |
PO Powerica | 2026-04-02 | Mainboard | -13.9% | +1.0% | -7.3% | 6.6% | Miss | |
SP Sai Parenterals | 2026-04-02 | Mainboard | +0.4% | +0.0% | +2.0% | 1.6% | Hit | |
AB Aeroplane Basmati Rice | 2026-04-02 | Mainboard | -10.2% | +0.9% | -5.7% | 4.5% | Hit | |
CM Central Mine Planning | 2026-03-30 | Mainboard | -8.9% | +5.2% | -7.0% | 1.9% | Hit | |
GC GSP Crop Science | 2026-03-24 | Mainboard | -4.1% | +0.0% | +2.5% | 6.6% | Miss | |
IN Innovision | 2026-03-23 | Mainboard | -10.9% | -4.8% | -9.9% | 1.0% | Hit | |
RS Rajputana Stainless | 2026-03-19 | Mainboard | -5% | +0.8% | +0.0% | 5.0% | Miss | |
SM SEDEMAC Mechatronics | 2026-03-11 | Mainboard | +7.1% | -1.5% | +13.5% | 6.4% | Miss | |
OE Omnitech Engineering | 2026-03-05 | Mainboard | -17.4% | +1.8% | -11.0% | 6.4% | Miss | |
PR PNGS Reva | 2026-03-04 | Mainboard | -5.5% | -0.5% | -2.8% | 2.7% | Hit |
Why Our AI Usually Beats Last-Day GMP
What last-day GMP actually is
Grey Market Premium is an unofficial, off-market quote of what traders are willing to pay for allotted shares before listing day. It reflects hype and liquidity - not fundamentals.
We use the last-day GMP(quoted on the IPO close date) because that's the number most retail investors actually see before they decide whether to apply. On hyped IPOs it tends to run way ahead of the real listing gain; on quiet mainboards it barely moves.
What our model uses
- Subscription velocity by category (retail / NII / QIB / anchor)
- Float size, peer-group multiple, and sector float absorption
- Fundamentals: revenue CAGR, PAT margin, ROE, debt/equity
- Issue structure: fresh issue vs OFS, anchor lock-in coverage
- Market regime: Nifty trend, VIX, and recent-listing performance
How we score accuracy
- Hit: prediction within 10% of actual listing gain
- Miss: prediction more than 10% off
- Direction: did we call gain vs loss correctly
- AI predictions are frozen on allotment day; last-day GMP is captured on the IPO close date
We're better, not perfect
The AI misses sometimes - it famously under-called Bajaj Housing Finance (+114% actual vs ~+95% pred) and Premier Energies (+120% vs ~+103%). Runaway hype listings are hard to model. But on average it lands far closer to reality than last-day GMP, and new listings are added to this dashboard the day after listing via our scraper pipeline - no cherry-picking.