Model Catalog
Four delivery tiers — Eco, Standard, Swift and Pro Ultra — one grounded quality bar.
Vedika exposes four delivery tiers for reasoning and AI Query calls. All four return the same source-grounded quality; they differ in latency and cost. You select a tier per request with the speed (and, for Standard, version) parameter.
#The four tiers
| Tier | Select with | Latency | Per query | Plans | Best for |
|---|---|---|---|---|---|
| Vedika Eco | speed:"eco" | Economy | from $0.01 | All plans | High-volume, cost-sensitive, simpler questions |
| Vedika Standard | speed:"standard" | ~120 tok/s · 20–30s | $0.01–$0.03 | All plans | The default — deep multi-factor synthesis |
| Vedika Standard 2.5 | version:"2.5" | ~120 tok/s | $0.01–$0.03 | All plans | Newer Standard variant on the standard path |
| Vedika Swift | speed:"fast" | ~1,800 tok/s · 1–3s | $0.02–$0.05 | Business+ | Lowest latency for chat UIs — same depth |
| Vedika Pro Ultra | speed:"standard" (deep profile) | ~120 tok/s | ~$0.03 | All plans | Maximum depth where quality > latency |
Note
The
speed parameter accepts eco, standard and fast. Vedika Pro Ultra is the deepest-synthesis profile of the standard path — the same speed:"standard" selector, tuned for maximum depth. Vedika Swift requires a Business or Enterprise key; a Starter or Professional key requesting speed:"fast" receives 403 PLAN_UPGRADE_REQUIRED.#Capability grid
| Capability | Eco | Standard | Swift | Pro Ultra |
|---|---|---|---|---|
Streaming (/query/stream) | Yes | Yes | Yes | Yes |
Structured output (responseFormat:"json") | Yes | Yes | Yes | Yes |
version (2 / 2.5) | — | Yes | — | — |
| Service tiers (flex/standard/priority) | — | Planned (2.5) | — | — |
Multi-turn conversationId | Yes | Yes | Yes | Yes |
| Prompt/response caching | Yes | Yes | Yes | Yes |
| Answer-cache hit price | $0.01 | $0.01 | $0.01 | $0.01 |
| Cached input tokens | Reduced rate (~60% of fresh) | Same as fresh input | Same as fresh input | Same as fresh input |
#Selecting a tier
Pass speed in the request body. For Standard you may also pin version to "2" (default) or "2.5":
{
"question": "...",
"birthDetails": { "...": "..." },
"speed": "eco"
}{
"question": "...",
"birthDetails": { "...": "..." },
"speed": "standard",
"version": "2.5"
}{
"question": "...",
"birthDetails": { "...": "..." },
"speed": "fast"
}Tip
Not sure which to use? See Choosing a model for a quick decision guide.