Mô hình
Nội dung tài liệu kỹ thuật hiện được cung cấp bằng tiếng Anh.
Open Mô hình to browse every model currently exposed through Uptech API.
What you see on the page
- API protocol filters —
chat.completions,responses,messages,generateContent,embeddings, and related shapes. - Model family filters — Anthropic, OpenAI, Google, DeepSeek, Moonshot, MiniMax, Alibaba, Zhipu AI, and more.
- Context window filters — 16K through 1M.
- Per-model cards — official slug, context size, input/output USD per 1M tokens, and tags such as
text,reasoning,vision,audio,long-context,cached.
Pricing is always USD per 1M tokens, with input and output billed separately. Cache hits use each model’s discounted cache rate when available.
How to choose a model ID
- Filter by protocol your SDK already speaks (most OpenAI SDK users start with
chat.completions). - Narrow by vendor or context window.
- Copy the official slug shown on the card.
- In API calls, use the exact slug shown in the catalog, for example
gpt-5.6-sol.
You can also list models programmatically:
curl https://api.uptech.vn/v1/models \
-H "Authorization: Bearer $APIGO_API_KEY"
From catalog to first call
- Copy a model ID from Models.
- Confirm you have balance under Workspace → Billing.
- Call
/v1/chat/completions(or the protocol that matches the filter you used). - Inspect Workspace → Call Logs to see which model actually answered (important when routing fallbacks are enabled).
Price changes
Uptech API mirrors upstream provider price changes. Adjustments are announced in-workspace with at least a 7-day transition window. Always re-check Models before locking a long-lived cost model into contracts.
