Reference
Models & pricing
All models, all prices, all tiers in one table.
§ 01
Tiers
Add tier to the request body to pick a performance level. Default is normal.
Prices in SEK per 1 million tokens (input / output):
- normal — 60 / 180 SEK. Standard real-time inference. Default for developer accounts.
- priority — 120 / 360 SEK (2× normal). Guaranteed first in queue. Never throttled. For time-critical applications.
- async — 30 / 90 SEK (½× normal). Requires
webhook_urland returns 202 with a job id. Result delivered via webhook within 12 hours — see the section below.
§ 02
Model
GET /v1/models — lists the active model (today glm-5.2).
- glm-5.2 — chat (GLM 5.2 FP8 på H200)
- bge-m3 — embeddings (1024-dim)
- bge-reranker-v2-m3 — rerank
- gliner-multi-pii — PII-detektering
§ 03
Pricing
API (per token, SEK per 1M, input / output). normal 60 / 180 · priority 120 / 360 · async 30 / 90. Prepaid wallet via card through Mollie, a cash invoice is sent automatically on each top-up.
| Modell | Enhet | Pris |
|---|---|---|
| Chat (GLM 5.2) | per token | 60 / 180 SEK per 1M tokens (in/ut, normal-tier) |
| Embeddings | per 1 M tokens in | 0.10 SEK / 1M input tokens |
| Rerank | per 1 000 queries | 0.50 SEK / 1,000 queries |
| PII-detektering | per 1 M tecken in | 0.20 SEK / 1M input characters |
Beta-fasen är gratis — vi spårar usage men debiterar inte. När vi aktiverar debitering meddelar vi 30 dagar i förväg.