Antibacterial screening software is sold on a monthly subscription, and on this site the published price runs from $149 a month for 250 screens up to $1,490 a month for 10,000. Divide one by the other and the useful number falls out: between $0.15 and $0.60 per compound screened against a full strain panel. That is the figure to carry into a budget conversation, because it is the one that compares directly against a plate.
Getting a comparable number out of anyone else in this category is the hard part. Almost no vendor selling computational screening into drug discovery publishes a price at all, so a team trying to build a screening budget ends up in three or four quote cycles before it can compare anything. Which vendors disclose anything, and what the few published ranges are, is laid out vendor by vendor on the drug discovery software cost page. This is a walk through the arithmetic, the four ways the market charges, and the specific things that move a quote up or down.
The four pricing models you will meet
They are not variations on one model. They bill different units, which is why quotes from two vendors are so often impossible to compare side by side.
| Model | Unit billed | Where you meet it | What it hides |
|---|---|---|---|
| Subscription per workspace | A monthly allowance of screens, plus seats | Prediction and triage software, including this one | The allowance. A tier that looks cheap can run out in week two. |
| Per compound quote | Compound, sometimes compound by strain | Contract labs running the assay for you | Setup, compound handling, minimum order size and repeat runs. |
| Per agent billing | One antimicrobial against one isolate | Clinical and veterinary reference labs | Panel arithmetic. Ten agents against one isolate is ten line items. |
| Free and academic | Nothing | Published classifiers and university servers | The terms. Retention and publication rights, and no support. |
The per agent model is worth seeing in the wild, because it is the clearest published evidence of how measured susceptibility is charged. ARUP Laboratories, a large US reference lab, states plainly that its individual MIC test bills per antibiotic tested, on a single bacterial isolate. Nothing about that is unreasonable, and it is exactly why a ten agent panel across six organisms turns into a number that needs approval.
Work out your own cost per compound
Ignore the sticker price and compute the unit. For subscription software the arithmetic is one line:
Monthly price divided by the monthly screen allowance, at the tier you will actually sit on.
On the tiers published here that gives $0.60 per compound on Lab, $0.25 on Discovery and $0.15 on Program, where one screen means one compound against a whole panel rather than one compound against one organism. The tier you will actually sit on is the part people get wrong. A team enumerating analogs in batches does not use 200 screens a month and then stop. It uses the allowance, hits the ceiling mid-series, and discovers the real price was the next tier up all along.
Two adjustments make the number honest. Add the seats you need, because a plan with two seats and a five person chemistry team is not the plan you bought. And check whether batch and programmatic access are in the tier or above it: on this site the antibacterial screening API is included from the Discovery tier and there is no per call surcharge, so a nightly job and a chemist in the browser draw on one allowance rather than two invoices.
What actually moves the quote
- Panel breadth. Six ESKAPE organisms or a custom panel of your own clinical isolates are different products. Custom panels almost always sit on a higher tier, everywhere.
- Batch ceiling. The difference between submitting fifty compounds and ten thousand is the difference between a tool and a pipeline component, and it is usually where the price step sits.
- Programmatic access. An API is commonly the line between a mid tier and an entry tier. If your read-out has to land in a registry, price the tier that has it and not the one you hoped would do.
- Seats and roles. Read-only seats matter more than people expect once a program has reviewers who should see results without spending the month.
- IP and retention terms. The cheapest option in the table is free, and the reason a company with an unpublished series cannot always use it is written in the terms, not the price list.
The comparison that decides the budget
Prediction is not competing with a plate. It is competing with the decision about which compounds get a plate, and that is where the money is. The consumable cost of broth microdilution is real but modest, and it is optimizable: a 2025 study in Antimicrobial Agents and Chemotherapy found a 15.1% cost reduction simply by moving MIC determination to 384 well plates at 30 µL, rising past 25.1% for expensive agents such as cefiderocol. Note what that tells you. If squeezing the volume saves a fifth of the cost, the consumables were never the expensive part.
The expensive parts are the compound and the slot. Synthesizing an analog that was never going to cross a Gram-negative outer membrane costs chemistry time you do not get back, and the assay slot it occupies was the one a better candidate needed. That is the arithmetic behind triage: the point of cutting a library before it reaches a plate is not to save on broth, it is to stop paying for the wrong compounds. At a quarter of a dollar per compound, a prediction pass over a 400 analog enumeration costs less than one repeat assay run.
It also helps to price the thing you are replacing honestly. If screening is outsourced, the invoice is only part of it: quoting, shipping compound, waiting on a queue and re-running failures all consume calendar time that does not appear on a purchase order. Small teams that track this properly tend to categorize those invoices into a monthly burn rather than treating each one as a one-off, and the pattern that emerges is usually a lot of small repeat runs that a triage step would have removed.
How much does antibacterial screening software cost?
Published subscription pricing in this category runs from roughly $150 a month for a single lab with a few hundred screens to around $1,500 a month for library scale runs with API access and custom isolates. Enterprise arrangements are quoted. Per compound, that lands between about $0.15 and $0.60 against a full strain panel, depending on the tier and how fully you use the allowance.
Is there free antibacterial screening software?
Yes, and some of it is very good. Published classifiers and university prediction servers cost nothing and are the right tool for mining a genome or filtering a large set of candidate sequences. The catch is rarely accuracy. It is the terms, the absence of per strain output, and the fact that many academic schemes reserve the right to retain submissions or publish them after an embargo, which is not workable for a series you intend to file on.
How much does MIC testing cost per compound?
Discovery stage MIC panels are quoted rather than listed, because the number depends on the organism count, the agent count, the replicate scheme and whether the lab has to handle your compound. The published reference points that do exist come from clinical and veterinary labs, which bill per antibiotic per isolate. Treat any single figure you see as a starting point for a quote, not a market price.
What should I ask before paying for screening software?
Five questions settle it. What is the monthly allowance and what happens when it runs out. Does one screen mean one compound or one compound against one organism. Are custom panels and your own isolates included at the tier you are pricing. Who owns the structures you submit, and are they used for training. And can the read-out leave the browser, as an export or through an API, without a second contract.
Where the money actually goes
A screening budget that works usually has three lines, in this order: a cheap computational pass that ranks everything, measured MICs on the short list, and a reserve for the repeat runs the short list will generate. The first line is the one that decides the size of the other two, which is why it is worth getting the unit cost right even though it is the smallest number on the page.
If you want the comparison in front of you, the breakdown of what screening tools actually return covers who does what in this market, the MIC prediction software page shows the read-out itself, and the plans are on the pricing page in dollars, with no quote cycle in front of them. If the budget question is whether a full computational chemistry suite is needed at all, the Schrodinger software cost write up covers how that licensing model prices.