Route the workload. Read the tradeoff.

Compare three declared model lanes on the units that shape a real agent budget: tokens, traffic, route share, price, and latency.

Assumption-led, not live data
The source names qualitative positioning but provides no current price sheet or SLA. Enter values you can defend; this lab computes the consequences.

Workload and route mix

Prices are USD per million tokens. Latency is an assumed average in milliseconds.

Requests per day

Input tokens per request

Output tokens per request

Sol

Route share %

Input price $/M

Output price $/M

Latency ms

Terra

Route share %

Input price $/M

Output price $/M

Latency ms

Luna

Route share %

Input price $/M

Output price $/M

Latency ms

Canonical result

No route yet

The useful budget view appears before export.

daily blended spend

--blended cost / request

--30-day spend

--weighted latency

--route share total

--all-Sol daily baseline

--spend reduction

How the blend movesEach lane cost = share x ((input tokens / 1M x input price) + (output tokens / 1M x output price)). Daily spend = blended request cost x requests/day. Latency is the share-weighted average.

Interpretation and limits

Use this to compare a scenario, not to certify a provider. Prices, traffic, tokens, and latency are declared assumptions. The model excludes quality, retries, cache hits, queueing, outages, and provider availability. A lower spend result does not establish equal capability.