What does your team need?
Choose the task and check your working conditions. You can see model options without completing a long form.
The shortlist is a starting point. Data handling and the complete working setup still need your organisation's checks.
What evidence is being compared?
Add details for your decision record Optional for browsing
Useful when your team is ready to document a decision. Leave these blank while exploring; no personal or confidential information is needed.
Try an illustrative workshop case
Presets fill the brief and priorities. They do not approve any organisation or deployment.
Your entries stay in memory until you save a record. Refreshing or closing this page loses unsaved changes.
What should the comparison focus on?
Use a starting mix, or keep the priorities already in your record. You do not need to set every weight to explore the options.
Fine-tune weights and record your reasoning Advanced · optional
Would moving from an unacceptable result to a good result make a large difference? A zero weight removes a category from scoring; mandatory requirements still apply.
More criteria: protection, resilience and organisational fit
How documentation contributes
Estimate cost for your workload
Advanced: score a matching raw cost or time measurement
Optional. Enter “good” and “unacceptable” values for the same workload and exactly the same unit as a raw observation. Values between them are mapped linearly. Blank or mismatched units leave the published scale unchanged. These are your value judgements, not benchmark findings.
Reasons are not needed to browse recommendations. Record them when finalising your team's decision.
Your shortlist, explained.
Explore the model options, their trade-offs and the checks before use. You can refine your priorities at any time.
Category sensitivity · ±20% one category at a time
Each active category is reduced and increased by 20%, keeping its internal allocation fixed, then the decision weights are renormalised. This checks those scenarios only; it is not a probability that an option is best.
Make the decision yours.
When you are ready, record your team's choice, its conditions and the next review. You can also save an unfinished draft.
Make it reviewable.
Print the complete record with weights, reasons, raw evidence, gate assessments, uncertainty and source URLs.
Save a record to keep your entries, or print the decision with its evidence and sources. This page keeps your entries in browser memory; they are not automatically stored or submitted.
Follow the evidence.
Exact model versions, test conditions and metric definitions matter. A benchmark result is evidence for the tested task.