Miatz playground

The Model-Landscape Board

Every model, ranked, priced, and told straight.

Coming to the Build-LabAIE-130ServerCross-Cuttinggen · dev
The Demystify signature

the same routing table Miatz's own inference layer uses internally to decide 'why did my run land on this model' is the public artifact — model selection never stays a vendor black box

How it works

What goes in, what comes out

What it does

A live-refreshed board pulls the Tier-C ranking feeds (rankings, real-usage leaderboards, cost trackers) into one filterable table/scatter. Learner plots any two axes — quality vs cost, quality vs latency, context window vs price — and watches the field rearrange; clicking a model opens its capability-tag card with a plain-language verdict.

You bring

pick two axes and a use-case preset; optionally star models into a personal shortlist

You get

interactive scatter/table that updates as feeds refresh, a plain-language verdict per model, an exportable shortlist

You control

axis pickers (cost/latency/quality/context/open-closed), provider filter, use-case presets (chatbot, coding agent, batch job)

Where it's used

FREE, top-of-funnel — /playgrounds gallery, linked from every lesson that names a model; the same read-only routing table doc 09 surfaces inside Cost & Latency Bench, just public and live

open vs closed weightsbenchmark literacycontext windowprice per tokencapability tiersrate limits

Concepts click when you open the machinery.

Three labs are already live and free — the same hands-on style this playground brings to its module.