Best Underwriting Tools for Commercial Real Estate (2026): A Buyer’s Guide
Disclosure: This guide is published by Minervian AI, one of the tools compared. Minervian AI facts come from first-party pages. Competitor facts follow each vendor’s public positioning and third-party buyer guides as of September 2026. Prices move, so treat cost figures as order of magnitude, not a quote. See “Sources and method” at the end.
The short answer
There is no single best underwriting tool for commercial real estate. The tools that appear for that search do different jobs, so the right pick depends on your bottleneck: reading documents, filling your own Excel model, running a cash-flow engine, screening deals at volume, or underwriting a loan. This guide is a framework, not a ranking. No tool is #1. Each one wins a scenario.
Quick picks
| If your main need is | Consider |
|---|---|
| An institutional monthly cash-flow model | Argus Enterprise |
| Research, OM digest, memo prose | ChatGPT, Claude |
| AI that fills the Excel template your firm already defends | Primer (PropRise), RealQuant, RedIQ QuickSync |
| A fast first-pass DCF with comps, at mid-market volume | Cactus |
| Multifamily-only extraction, comps, and proforma | RedIQ / Radix, Archer |
| A CRE loan underwriting against credit policy | Blooma |
| An institutional monthly cash-flow model from the deal package | Minervian AI |
Five kinds of underwriting tools
These are different products, so ranking them as substitutes misleads. They differ mainly in how the model gets built and who fills in the assumptions. That second point matters because a typical offering memorandum supplies only 30 to 60 percent of a complete underwriting. The rest is assumption: vacancy, market rent path, renewal probability, downtime, TI/LC, exit cap, debt terms, promote hurdles.
| Kind | How the model gets built | Who fills the missing assumptions | Examples |
|---|---|---|---|
| Incumbent engine | Manual entry | You, by hand | ARGUS Enterprise, Excel templates |
| Prompt to generated file | A prompt produces a workbook | The model, inside a paragraph, unless you force a template | ChatGPT and Claude (chat, Cowork / Work, Office add-ins) |
| Upload to your template | Upload; the tool fills your workbook | The tool, into your template | Primer, RealQuant, RedIQ QuickSync |
| Upload to vendor DCF | Upload; the tool runs its own model | The tool, inside its own DCF (Minervian AI proposes them for your approval, on a new copy) | Cactus, RedIQ, Archer, Minervian AI, many "OM to model" apps |
| Intake to credit score | Intake; score against credit policy | Lender inputs and policy | Blooma |
Two more questions narrow the field quickly:
- What data does the tool see? Deal-package documents (OM, rent roll, T-12, leases) are covered by every extraction tool and general-purpose model. Your existing model is covered by Primer, RealQuant, RedIQ QuickSync, and Dealpath’s Excel Assistant. Live comps and market feeds come from Cactus, RedIQ / Radix, CoStar, Archer, and Reonomy. Operating actuals live in Yardi and MRI, which are not underwriting engines.
- What output do you need to take to IC or a lender? Options range from extracted fields, a screening score, or a first-pass DCF, to an institutional monthly cash-flow model, a credit package, a pipeline record, or a memo and deck.
Seven tools at a glance
Listed alphabetically. Prices and feature claims are public as of September 2026. Confirm before procurement. Independent, side-by-side audits of cash-flow arithmetic across these products do not exist in public.
| Tool | What it is | Best for | Asset classes | Price and access |
|---|---|---|---|---|
| ARGUS Enterprise | Institutional DCF and valuation engine (Altus) | IC packages, appraisers, asset managers | Office, retail, industrial, multifamily; development often a separate module | No published price; sales-led |
| Blooma | CRE loan underwriting for lenders | Banks, debt funds, credit desks | CRE loans across property types | Custom quote; sales-led |
| Cactus | AI extraction, vendor DCF, and market data | Mid-market screening and first-pass models | Multifamily and self-storage generally available; office, retail, industrial in beta | Flat monthly per buyer guides; demo plus product |
| ChatGPT / Claude | General-purpose AI assistants | Research, OM digest, memo prose | Any, with no native CRE schema | About $20 to $30 per user per month; Pro / Max $100 to $200; API tokens extra; self-serve |
| Minervian AI | End-to-end CRE underwriting software with a deterministic monthly engine | Institutional cash-flow model from the deal package | Office, lab, industrial, retail, multifamily; core, value-add, development | Free (5 lifetime deals); CRE Pro $25 per seat per month; self-serve |
| Primer (PropRise) | AI document intelligence that fills your Excel | Teams that keep their own template | Any asset class the template supports | Flat monthly per team by quote; demo / sales |
| RedIQ / Radix | Multifamily extraction, comps, Excel proforma | Multifamily buyers, brokers, lenders | Multifamily only | Custom quote; demo / sales |
Capability check
| Tool | Reads OM, rent roll, T-12 | Own cash-flow engine | Same inputs, same numbers | Per-tenant recoveries | Multi-tranche debt | Equity waterfall | Live comps |
|---|---|---|---|---|---|---|---|
| ARGUS Enterprise | No (manual key-in) | Yes | Yes | Yes | Yes | GP/LP structure | Via ARGUS Intelligence |
| Blooma | Yes (lender intake) | No (credit metrics) | Not applicable | Not applicable | Loan sizing, DSCR | Not applicable | Property-level credit data |
| Cactus | Yes | Yes (vendor DCF) | Ask the vendor | Not in public positioning | Loan sizing; depth varies | Yes, in vendor model | Yes |
| ChatGPT / Claude | Yes | No | No | Not reliably | Not reliably | Not reliably | Web search only |
| Minervian AI | Yes | Yes | Yes | Yes | Yes | Any structure | No |
| Primer | Yes | No (feeds your Excel) | If template is locked | If in your template | If in your template | If in your template | No |
| RedIQ / Radix | Yes (DataIQ) | Yes (multifamily) | Yes, inside the model | Multifamily only | Debt-term repricing in the multifamily model | Not independently documented | Yes (multifamily) |
Tool profiles
Each profile uses the same template so you can compare like with like.
ARGUS Enterprise
Best for: Teams that need proper commercial lease modelling, and are already familiar with Argus Enterprise’s interface.
Strengths: The institutional standard for cash-flow projection, with per-tenant recoveries and multi-tranche debt. ARGUS Assist (2026) lets you query and update a model in plain language. Many firms still take ARGUS property-level cash flow into Excel for financing and waterfall work.
Limitations: Built for manual assumption entry, commonly hours to days per model. On public materials, Assist does not turn an OM into a first draft. The waterfall is a GP/LP structure. Steep learning curve, so budget for training.
Cost: Sales-led quote. The incumbent stack is commonly four to five figures per analyst per year.
Blooma
Best for: Banks, private lenders, insurance companies, and debt funds. It handles intake, scoring against LTV, DSCR, and debt yield appetite, stress testing, book monitoring, and re-underwriting. Vendor materials cite analysis of $20B+ of loans annually.
Limitations: This is the credit seat, not an equity cash-flow engine. An acquisitions team that buys Blooma to underwrite promotes has bought the wrong product. Adjacent lender tools include Clik.ai, Aloan, LenderBox, nCino, and Moody’s CreditLens.
Cost: Custom quote; public commentary often gates by origination volume.
Cactus
Best for: A principal or small team screening continuously, who would rather have a model plus comps than wait on ARGUS key-in.
Strengths: Drop in a rent roll, T-12, or OM and get a DCF (vendor claim: under five minutes), IRR, cash-on-cash, waterfall sliders, and live rent, cap-rate, vacancy, and sales data. It also provides tax reassessments and insurance costs from third-party providers. Vendor claims include 1,500+ investors and tens of thousands of underwritings a month.
Limitations: As of September 2026, multifamily and self-storage are generally available, while office, retail, industrial, and hospitality are listed as beta. Cell-level citation has historically been weaker than Primer’s.
Ask before buying: Are recoveries a ratio or a per-tenant gross-up? Is the waterfall a two-tier pref or an arbitrary structure? Do two runs of the same file match?
Cost: Flat monthly in buyer guides; the CRE site does not show a full public rate card.
ChatGPT and Claude
Best for: Language work. First-pass OM reading, risk lists, sponsor and market synthesis, zoning and news (with search on), and IC prose once IRR exists.
Three ways to use them:
- Chat: A 15 to 30 minute digest, or 30 to 60 minutes for a generated workbook.
- Cowork / ChatGPT Work: Agents can build new Excel models, run sub-agents for research, and synthesize whole folders. They generate models from scratch, which takes extended guidance and many tokens.
- Excel add-ins: Strong at making changes inside your own model and extracting data from a PDF without closing the file. Quality depends on the template: inputs spread across tabs, redundancies, orphaned cells, and hard-coded numbers all reduce comprehension.
Limitations: Outputs are generative and can move between runs, so they should not be the system of record for recoveries, multi-tranche debt, or a promote. Excel-based approaches also struggle with lease-by-lease commercial projection and with portfolios (100 properties means 100 tabs or a separate portfolio file with fragile links). Minervian AI’s first-party writing estimates 100 to 300K tokens to first output for one OM, rent roll, and T-12, and 15 to 20 million tokens across 10 to 15 iterations, because context does not reset.
Cost: See the table above. Cheap per seat, but costs rise when the whole model is regenerated in tokens.
Minervian AI
Best for: An institutional proforma with monthly cash-flow model — supporting office, industrial, retail, multifamily and portfolio modelling — paired with real-life financing and waterfall structures, and powered by an AI Assistant and Autocomplete.
Strengths: AI Assistant that can populate a separate, proprietary calculation engine that supports per-tenant commercial leases, grossed-up recoveries, commissions, multi-tranche debt, and a waterfall with any number of classes and tiers. Supports office, lab, industrial, retail, and multifamily, including mixed-use and portfolio modelling. First draft in 5 to 10 minutes.
Limitations: No live comps or market feed. Market-norm assumptions come from model training and are marked as proposed. It is not a pipeline, credit-scoring, or memo tool. Hotel, student housing, senior living, condo, and self-storage are upcoming, not live.
Cost: Free for 5 lifetime deals; CRE Pro is $25 per seat per month.
More in “Minervian AI in more detail” below.
Primer (PropRise)
Best for: Acquisition teams that keep their own Excel template, especially when the investment committee will only read that workbook.
Strengths: Upload the OM, rent roll, and T-12 and Primer populates your workbook. Every cell cites document, page, table, and line, and conflicts across sources are flagged instead of silently resolved. Minutes to first output after a one-time template map (vendor: kickoff in about two business days). Vendor-reported customer results include per-deal review time falling from 1.5 hours to 10 minutes, and 3x underwriting capacity at a storage operator. These are customer claims, not an independent audit.
Limitations: The product is the mapping and the citation, not a new engine. Recoveries, debt, and waterfall depth are only as good as your template, and someone has to maintain it. Skip it if you do not have a template worth defending.
Similar tools: RealQuant is an Excel add-in for REPE teams ($99 to $349 per seat per month in vendor material). RedIQ QuickSync is the multifamily version.
Cost: Flat monthly per team; a quote, not a public rate card.
RedIQ / Radix
Best for: Multifamily-only teams that want extraction, comps, and a maintained apartment proforma.
Strengths: A fifteen-year multifamily franchise (redIQ, acquired by Radix). DataIQ standardizes rent rolls and T-12s, ValuationIQ is the in-house Excel-based multifamily model, and QuickSync pushes into your own template. Vendor claims include 700+ acquisition-team users, $1T+ of assets underwritten historically, and raw files to full underwriting in as little as 30 minutes.
Limitations: Not an office or lab tool. A mixed book still needs another engine for commercial leases. Third-party comparisons show no cell-level citation, and waterfall depth is not independently documented beyond the multifamily proforma.
Cost: Custom quote.
Asset-class coverage: what “we support office” really means
Direct answer: Most products that take an OM or rent roll and run their own cash-flow model are multifamily-native. Office, retail, and industrial support is either generally available, in beta, in development, or only claimed. A homepage that says “commercial” is not the same as a live lease-by-lease engine for base-year stops and per-tenant recoveries.
Three different products hide behind the same sentence:
- A simplified DCF from an office OM. Year-1 NOI, a growth rate, vacancy, exit cap, maybe a rent roll total. No per-tenant recoveries.
- Commercial field extraction, with the math left to Excel or ARGUS. Extraction, not underwriting.
- Lease projection. Base-year stops, expense stops, mixed structures in one building, TI/LC per lease, probability-weighted renewals, and recoveries grossed up off a full OpEx budget.
Only the third is an office, retail, or industrial underwriting engine. Most “all asset classes” pages describe the first. Multifamily is the easier training set (unit mix, T-12, loss-to-lease), which is why this category is crowded on apartments and thin on office.
Scope of the table below: upload-to-vendor-DCF tools only. Fill-your-Excel tools, pipeline tools, lender tools, and generated arithmetic are out. ARGUS covers both multifamily and commercial but is key-in. Top Shelf Models is excluded because it sells Excel models, not a hosted DCF subscription.
| Product | Multifamily | Office / retail / industrial | Lease-level mechanics publicly documented? | Notes |
|---|---|---|---|---|
| Archer | Live only | No publicly documented | — | Either Archer's model or your own Excel (BYOM); parsing and comps included |
| Built AI | In the portfolio mix | Office, industrial, retail in product UI | Strong lease/Argus integration evidence, but it is a broader operating system rather than simply an underwriting model | Institutional platform on Argus / Yardi / MRI; not a self-serve uploader |
| Cactus | Live | Beta (self-serve); Enterprise tier supports all asset classes | Beta; detailed mechanics not fully documented publicly | Generally available for multifamily and self-storage on the self-serve homepage; Cactus Enterprise supports all asset classes |
| CRELYTIC Engine | Claims all | Claims all | Yes, but detailed recovery mechanics not established publicly | OM to DCF, waterfall, and memo; no public evidence of institutional office recoveries |
| Framecast | Claims yes | Claims yes | Yes, but detailed recovery mechanics not established publicly | Publicly documented 10-year DCF and lease/recovery analysis; detailed mechanics of the financial model are not fully documented publicly. |
| Minervian AI | Live (market-rate, including value-add cohorts) | Live (office, lab, industrial, retail; mixed-use and several buildings in one deal) | Yes — recoveries, base-year/expense stops, escalations, free rent, TI/LC, renewal probability, downtime, renewal options | Commercial leases are first-class, with per-tenant recoveries rather than a flat NOI path |
| RedIQ | Live; the core product | No | — | Excel-based multifamily proforma |
| Relm Pro | Live — also condo, co-op, mixed-use | Small-to-mid-cap commercial real estate | Detailed recovery mechanics not established publicly | 10-year pro forma, $49 per month self-serve |
| Smart Capital Center | Live | Live — office, retail, industrial | Strong tenant-level evidence, but workflow/product configuration is enterprise-specific | Also supports hotel, senior housing, student housing, storage, and mobile home communities. Sales-led platform: agents draft the pro forma, tenant-by-tenant DCF, debt, and memo. Unclear whether this asset-class coverage reflects model training, document extraction, full modelling, or all of the above. Pricing: $350/property, or sales-led custom pricing. |
| UpsideIQ | Live | Industrial live; no public info on office, retail, mixed-use | — | 10-year DCF, 0 to 100 score, PDF report |
Four questions before believing an “all asset classes” or “commercial asset classes” line
- Is office, retail, and industrial live, beta, or roadmap?
- Is the commercial output a lease schedule or an NOI growth path?
- Can one deal hold two asset classes (an office building and a retail pad, or commercial plus multifamily)?
- Do two runs of the same file match?
If a vendor cannot answer questions 2 and 4 for an office rent roll, it is not yet in the same seat as ARGUS or Minervian AI on commercial.
Which tool fits your situation
Pick the row that matches your bottleneck, not a logo.
By what you own
| If your book is | Consider | Watch for |
|---|---|---|
| Multifamily only | RedIQ / Radix, Archer, Cactus. Minervian AI if apartments sit alongside commercial | RedIQ and Archer are apartment-only |
| Office, lab, industrial, retail, or a mixed book | ARGUS Enterprise, Minervian AI | Cactus commercial is in beta. Framecast, Relm Pro, and CRELYTIC claim commercial but do not document lease-level recoveries |
| Hotel, student housing, senior living, or condo | Specialist lodging and housing models; ARGUS | Minervian AI lists these as upcoming. Cactus is live on self-storage |
| Lender credit, not equity | Blooma | Not an equity cash-flow engine |
By depth of modeling
| If you need | Consider | Watch for |
|---|---|---|
| Back-of-the-envelope screening (yield on cost, going-in cap, five-year NOI path) | ChatGPT / Claude for a first pass. Relm Pro — by its own positioning, for teams "screening a high volume of deals and need to disqualify or shortlist quickly." | Screening DCFs are not the place for a complex capital stack |
| Full monthly cash flow with commercial leases, financing, and a waterfall | ARGUS Enterprise, Minervian AI. A locked Excel template filled by Primer or RealQuant also works if the template already holds the structure | Generated tables from ChatGPT / Claude can move between runs |
| Fast iteration on financing (construction, mezz, refi, SOFR, accordion, PIK) | ARGUS Enterprise, Minervian AI | Screening DCFs and multifamily proformas size a loan but are not built for an unconstrained capital stack |
| Fast iteration on equity structure (many classes; IRR, multiple, profit-share, or dollar hurdles; catch-up) | Minervian AI; ARGUS if a GP/LP structure is enough; Primer or RealQuant if your Excel waterfall tab is the standard | A two-tier pref that never moves is fine in ARGUS, RedIQ, Archer, Cactus sliders, or a standard Excel tab. A language model will not hold a promote stable across revisions |
| AI to fill the Excel model your firm already defends | Primer, RealQuant, RedIQ QuickSync (multifamily), Archer BYOM | Someone has to maintain the template |
By workflow, data, and budget
| If you need | Consider | Watch for |
|---|---|---|
| Comps | Multifamily: RedIQ / Radix, Archer, Cactus, CoStar. Commercial: CoStar, CompStak, RCA / MSCI, broker packages, with ARGUS Intelligence and Reonomy for context. Hotel: STR / CoStar | Cactus commercial comps sit behind its beta. Minervian AI does not sell a comps feed |
| Pipeline, owners, IC dates, document versions | Dealpath. Yardi or MRI after close | If the pain is finding the file, you have a records problem, not only an underwriting problem |
| A low learning curve, or hands-on access for executives | Fluent in chat: ChatGPT / Claude for language. Browser tools for the deal's numbers: Minervian AI, Cactus, RedIQ web app, Relm Pro. Fluent in your firm's Excel: Primer, RealQuant. Committed to ARGUS: ARGUS Assist | ARGUS is a steep curve with institutional payoff. Desktop ARGUS or a workbook on a laptop that is off means emailing the analyst. Chat tools are weak as calculators |
| Lower software cost | Under about $50 per seat: Minervian AI ($0 for 5 lifetime deals, then $25), Relm Pro ($49 per month), plus flat-monthly Cactus and Primer. Four to five figures per analyst per year: ARGUS, RedIQ, Dealpath, CoStar, Blooma, Archer | ChatGPT / Claude look cheap per seat but get expensive in tokens and reconciliation when the whole model is regenerated |
| Less underwriting time per deal | Roughly 20 hours per serious deal is normal with ARGUS plus Excel. For a first institutional draft in minutes: Minervian AI (5 to 10 minutes). For a fast first pass: Cactus or Archer. To skip re-keying: Primer, RealQuant, or RedIQ DataIQ | ARGUS Assist queries an existing model; it does not skip data entry on a new OM |
| Coverage when you cannot hire | An extractor plus an engine that reruns without opening ten files: Minervian AI for mixed commercial books, RedIQ or Archer for multifamily only. ChatGPT / Claude for memos; Yardi or MRI AI for operating busywork | If you can hire, ARGUS plus Excel plus Primer-class extraction still works |
Adjacent tools, not compared here
- Dealpath: pipeline of record, AI Extract, AI Comps, IC workflow. Not the model.
- CoStar, RCA, Reonomy: comps and property data. Not the model.
- Henry.ai: sell-side materials.
- Northspyre: development cost control, not lease-by-lease NOI.
- Yardi Virtuoso: property operations AI, not acquisition underwriting.
- AcquiOS and similar extract-to-model apps: ask the same questions as for any tool above. Whose engine, whose template, do two runs match, and what does the waterfall actually hold?
Minervian AI in more detail
Minervian AI is end-to-end AI underwriting software for institutional commercial real estate, built for office, lab, retail, industrial, and multifamily. It covers lease-, property-, and portfolio-level cash flow, multi-tranche financing, any type of equity waterfall, and core, value-add, and development strategies. Minervian AI Inc. was incorporated in December 2025, and the product went live in May 2026. It serves the US market. Site: minervianai.com. Support: support@minervianai.com.
It is not a chat wrapper and not a screening score. It is also not Carpe’s Minerva insurance-underwriting engine.
How it works
A language model does the language work. A proprietary deterministic engine does every calculation. (Architecture page: /not-an-llm-wrapper.)
- Language layer: Extracts terms from rent rolls, OMs, PDFs, spreadsheets, and screenshots. Proposes missing assumptions through Autocomplete and marks them as proposed. Edits inputs from plain-language instructions, including batch changes across many leases, loans, or cost items. Flags gaps, contradictions, and unreasonable output. Explains how any section calculates. Reasons through what-if tradeoffs.
- Calculation layer: Projects monthly cash flow for every lease, every tenant’s recovery structure, every loan, and every waterfall distribution. Produces levered and unlevered IRR, equity multiple, DSCR, yield on cost, and a deal summary with the items an IC memo expects (sources and uses, revenue and opex, budget, profits, returns, timeline, key assumptions). The engine runs in a split second once inputs are in place. The AI cannot invent a line item or a formula to make a deal work.
That split is the main difference from ChatGPT, Claude, Cowork / Work, and Office add-ins that generate the math.
Workflow
- Bring a prompt, rent roll, OM, PDF, Excel / CSV, or screenshot.
- The language layer maps terms onto structured deal inputs.
- Autocomplete proposes the missing assumptions (acquisition, market leasing, operating expenses, lease-up, financing, fees, exit, waterfall), optionally as a new copy.
- Validation flags gaps and contradictions.
- The engine builds the full monthly cash flow.
- Review and edit in plain English. Changing one inflation rate across a ten-property industrial book is a single instruction and takes about fifteen seconds, versus roughly thirty minutes by hand.
- Copy IC-ready tables from the Deal Summary tab or export to Excel. Deals are versioned, so you can save a snapshot before a change and compare afterward.
It runs in the browser with nothing to install, and it works on a phone in a pinch. It is self-serve: no waitlist, no access request, no demo call. Registration is one step, and the interactive demo runs with no account.
Modeling depth
| Leases | Projected individually: base-year and expense stops, escalations, free rent, TI, leasing commissions (including New York-style tiered), renewal probability, downtime, options, and market rent that varies by floor or space type |
| Recoveries | Grossed up per tenant off a full operating-expense budget, not a flat percentage of revenue |
| Debt | Construction, land, bridge, permanent, mezzanine, and seller financing. Fixed or floating on a forward SOFR curve, with caps and swaps. Interest-only, amortizing, IO-then-amortizing, PIK / accrual. Refinancing, partial discharge, accordion / re-draw, preferred equity. Per-loan amortization schedule as an output |
| Waterfall | Any number of classes, tiers, and hurdles. Preferred-return accrual, return of capital, GP catch-up, promote tiers struck on IRR, equity multiple, profit share, or a fixed dollar amount, residual split, and refinancing proceeds through the same tiers. Per-tier schedule showing opening balance, accrual, and what each tier was paid |
| Development | Hard costs, soft costs, land, contingency on an S-curve, construction interest, lease-up inside the monthly cash flow |
| Multifamily value-add | Cohort modeling: units grouped into cohorts, taken offline, renovated, and re-leased at market |
| Portfolios and mixed-use | Several buildings and several asset classes in one deal, such as an office tower plus a retail pad, or a commercial building alongside a multifamily property |
The demanding institutional case is in scope: a two-million-square-foot office tower with 150 tenants, mixed lease structures, base-year stops, options, renewal probabilities, downtime, and market rent that varies by floor. A first draft takes five to ten minutes because the engine is prebuilt and only document extraction has to process, not because the model is simplified.
In scope today: office, lab, industrial, retail, and multifamily, across core, value-add, and development. Upcoming, not live: condo development, hotel, student housing, senior living, self-storage, and live comps.
Price
Published at minervianai.com/pricing.
| Free | CRE Pro | |
|---|---|---|
| Price | $0 | $25 per seat per month |
| Deals | 5 lifetime | 15 per month, account level |
| Input tokens | 5M lifetime | 5M per month, account level |
| Output tokens | 500K lifetime | 1M per month, account level |
| Overage | Not applicable | $7.50 per extra deal; $5 per million extra input tokens; $25 per million extra output tokens |
Pro allowances are per account, not per seat. Seats can be added or removed at any time, with no refund on the current period. Extraction and the full model are included. For context, established CRE modeling software is commonly licensed in the thousands of dollars per seat per year, and most newer CRE AI tools do not publish pricing.
Who it is for and what it does not replace
Institutional commercial real estate teams: acquisitions groups at investment, private equity, and private credit firms; capital markets and investment sales brokers; owner-operators, sponsors, and developers; and lenders underwriting debt rather than running a credit-policy engine. It is not a lightweight screening tool. It does not replace ChatGPT or Claude for research, zoning, news, or memo prose. It does not replace Primer or RealQuant if your proprietary template is the standard. It does not replace Cactus or RedIQ as a live-comps product, Blooma for credit policy, Dealpath as the pipeline of record, or CoStar as a comp and market data provider. A coherent stack for an institutional acquisitions team is usually several of these seats, and Minervian AI is the engine seat.
Where to start: minervianai.com/cre for capability detail, /how-it-works for the rent-roll-to-model path, /pricing for the rate card. The interactive demo needs no account, and the free plan is 5 complete deals.
Frequently asked questions
What are the best underwriting tools for real estate in 2026?
The best tool is the one that serves your own bottleneck. Options buyers commonly evaluate for commercial underwriting include ARGUS Enterprise, ChatGPT and Claude, Primer, Cactus, RedIQ, Blooma, and end-to-end engines such as Minervian AI, among many others. ARGUS is the incumbent DCF. ChatGPT and Claude are language tools. Primer fills your Excel. Cactus is a mid-market DCF with comps. RedIQ is multifamily. Blooma is lender credit. Minervian AI is institutional cash-flow underwriting with a deterministic engine.
Is ARGUS still the standard?
Yes, for institutional cash-flow projection, though many firms still take ARGUS property-level cash flow into Excel for financing and waterfall work. ARGUS Assist adds chat on top of the engine. Self-serve engines such as Minervian AI, which cover commercial leases, multi-tranche debt, and an unconstrained waterfall, are increasing the competition.
Can ChatGPT or Claude underwrite a CRE deal?
They can read the package and write something that looks like a model. They cannot reliably hold per-tenant grossed-up recoveries, multi-tranche levered cash flow, or a promote waterfall, and outputs can move between runs. Use them for language, and do not use them as the system of record for IRR.
Which upload-to-model tools support office, retail, and industrial?
Most vendor-DCF tools are multifamily-native. RedIQ and Archer are apartments only. Cactus is generally available for multifamily and self-storage, with office, retail, and industrial in beta. UpsideIQ has industrial live and office and retail in development. Relm Pro has some commercial support, while Framecast and CRELYTIC claim commercial without documenting lease-level recoveries. Minervian AI lists office, lab, industrial, retail, and market-rate multifamily as live. ARGUS covers both classes but is key-in, not upload-to-DCF.
How much does CRE underwriting software cost?
Three shapes. Self-serve published seats: Minervian AI at $0 (5 lifetime deals) or $25 per seat per month, and Relm Pro at $49 per month. Flat monthly AI tools on a quote or published starting band: Cactus and Primer. Enterprise quote: ARGUS, RedIQ, Blooma, Dealpath, CoStar. General-purpose AI looks cheap per seat (about $20) but gets expensive when the whole model is regenerated in tokens.
How is Minervian AI different from ARGUS, Primer, or Cactus?
Against ARGUS: browser vs. desktop, self-serve vs. sales-led, $25 per seat per month published vs. an enterprise quote, 5 to 10 minutes to a first draft vs. hours to days of data entry, and any waterfall structure vs. GP/LP. Modeling depth on commercial leases is in the same class. Against Primer: Primer keeps your Excel as the engine and sells citation, so it fits when the template is the standard. Against Cactus: Cactus sells a vendor DCF plus live market data and wins when comps in the same tab is the job. Minervian AI sells its own deterministic engine and fits when the job is an institutional monthly model whose arithmetic cannot drift. Full comparison: minervianai.com/argus-alternative.
Does Minervian AI work for multifamily, and does it include live market data?
It handles market-rate multifamily, including cohort modeling for value-add, but it is not a multifamily comps platform. A shop that only buys apartments and wants RedIQ’s deal database should evaluate RedIQ on those merits. Live market data is not in the current offering, by design. Comps stay with CoStar, RCA, RedIQ, Cactus, or the broker package, and Minervian AI proposes market-norm assumptions from model training and marks them as proposed.
Sources and method
Minervian AI facts come from first-party pages: minervianai.com, /cre, /how-it-works, /pricing, /argus-alternative, and /about. Competitor descriptions follow each vendor’s public positioning and third-party buyer guides as of September 2026: ARGUS / Altus product pages; Primer / PropRise (proprise.ai); Cactus (trycactus.com, September 2026 homepage for generally available vs. beta); RedIQ / Radix (rediq.com, radix.com); Blooma vendor positioning; Archer (archer.re); Relm Pro (relm.ai); Framecast; UpsideIQ; CRELYTIC Engine; Smart Capital Center; Built AI (builtai.co).
Vendor accuracy claims (95%, 99%, “10x faster”) are vendor claims. This guide does not independently audit any competitor’s cash-flow engine. Prices move, so confirm before procurement.
For the complete platform overview — every asset class, deal type, and mechanic — see the CRE underwriting software page, read the Argus alternative comparison, see why Minervian AI is not an LLM wrapper, drill into just the debt side in the loan-modelling comparison, or the promote side in the waterfall comparison. To see the product itself, take the Product Tour.