Modellwerk
tensor grouped query head map
Map query heads to shared key-value heads and compute declared KV tensor element savings.
POST https://modell.halowerk.com/v1/tensor-grouped-query-head-mapInputs
| Field | Type | Required | Description |
|---|---|---|---|
| query_heads | integer | Yes | |
| kv_heads | integer | Yes | |
| head_dim | integer | Yes | |
| tokens | integer | Yes | |
| layers | integer | Yes |
Show full input schema
{
"type": "object",
"properties": {
"query_heads": {
"type": "integer",
"minimum": 1,
"maximum": 256
},
"kv_heads": {
"type": "integer",
"minimum": 1,
"maximum": 256
},
"head_dim": {
"type": "integer",
"minimum": 1,
"maximum": 1024
},
"tokens": {
"type": "integer",
"minimum": 1,
"maximum": 1000000
},
"layers": {
"type": "integer",
"minimum": 1,
"maximum": 256
}
},
"required": [
"query_heads",
"kv_heads",
"head_dim",
"tokens",
"layers"
],
"additionalProperties": false
}Published input example
{
"query_heads": 8,
"kv_heads": 2,
"head_dim": 64,
"tokens": 100,
"layers": 2
}Result & documentation
Machine-readable documentation of this hub
Source of this page: the published HALOWERK catalog. Updated: 2026-11-18 UTC.
More endpoints in this hub
- Modellwerk0.003 USDC
chat completions
POSThttps://modell.halowerk.com/v1/chat-completionsRuns openai/gpt-oss-20b via GroqCloud. Messages are sent to Groq Inc., USA; payment metadata is not forwarded. At most 8 messages, 8192 UTF-8 bytes overall and 1024 completion tokens including reasoning. A length finish_reason exposes trunc
- Modellwerk0.002 USDC
context plan
POSThttps://modell.halowerk.com/v1/context-planPlans the division of a long input for a given model. It reserves room for the system prompt, the expected answer and a safety margin, then works out how many chunks are needed and where to cut, preferring paragraph boundaries over line and
- Modellwerk0.002 USDC
cost estimate
POSThttps://modell.halowerk.com/v1/cost-estimateWorks out what a call would cost. Give either the text itself or a token count, plus the expected output length, and get input, output, cache-read and cache-write cost per model with the total. Token counts derived from text are estimates f