ModelSpec · model decision prototype

Task

Tokens per task

in out

Rank by (weights sum to 1)

{{ w.label }}{{ w.valText }}

Reading your task…

Read from your task:

“{{ t.word }}” → {{ t.note }}

Edit condition{{ edit.title }}

Narrowing, in the order you set conditions{{ funnelSummary }}

Next questions, most narrowing first

No further question would narrow the field.

Couldn't load snapshot {{ snapshot }}.

The catalogue service didn't respond in 8 s. Your spec is kept in the link, so nothing is lost. Results below are hidden rather than shown stale.

Trade-off canvas

x y

Qualifies · independent Qualifies · lab-reported only May qualify Excluded Interval

{{ emptyTitle }} {{ emptyBody }}

{{ n.name }} · {{ n.why }}

↑ {{ yLabel }}{{ xLabel }} →

Not plotted, value unknown on this axis: {{ unplotted }}

Shortlist across trade-offs{{ shortSummary }}

May qualify · {{ mayCount }}Passes everything we know

None. Every remaining model is known on every condition.

Show {{ exclCount }} excluded, with reasons


{{ r.rank }}	{{ r.name }} · {{ r.lab }}	{{ r.via }}	{{ r.y }} {{ r.ci }} {{ r.provMark }}	{{ r.cost }}	{{ r.pin }}	{{ r.pout }}	{{ r.ttft }}	{{ r.tps }}	{{ r.ctx }}	{{ r.weights }}	{{ r.status }} {{ r.reason }}

Eliminations by condition

{{ g.label }} · {{ g.n }}

{{ d.status }}{{ d.statusWhy }}

Provisional. {{ d.provText }}

Retired, in the live archive. {{ d.retText }}

Evidence too thin to separate these. {{ d.insepText }}

What each weight contributes

{{ c.label }}{{ c.value }}

Outlined track = the weight you set. Filled = how much of it this model earns against the others that qualify. Conditions never add to this; they only filter.

Tipping point {{ tip.text }}

$ per task weight 0{{ tip.cur }}1

Evidence

Facts

What the lab didn't report {{ d.missing }}

Against your conditions · {{ d.offName }}

What each condition costs you

{{ c.label }}{{ c.pts }}

Near misses

{{ n.name }} would qualify if you relaxed “{{ n.cond }}”: {{ n.why }}

No model is one condition away.

Why not…

{{ wn.head }}

Offerings · same model, different terms{{ d.derivation }}

Provider	Regions	Input $/1M	Output $/1M	$ per task	First token	Output tok/s	Retention	Against your conditions
{{ o.provider }} {{ o.used }}	{{ o.regions }}	{{ o.pin }}	{{ o.pout }}	{{ o.cost }}	{{ o.ttft }}	{{ o.tps }}	{{ o.ret }}	{{ o.status }} {{ o.why }}

Select a model on the canvas, in the shortlist or in the table to see why it lands where it does.

Loading snapshot {{ snapshot }} · {{ stats.models }} models, {{ stats.offerings }} offerings

Which AI model fits your task, under your constraints, and why.

{{ d.name }}