Language Models
Use one Chat Completions interface for conversation, writing, summarization, translation, and classification.
Call language models from different providers through one OpenAI-compatible API for question answering, content generation, summaries, translation, classification, and information extraction.
Use cases
Question answering
Product support, customer-service assistance, internal knowledge, and open-domain conversations.
Content generation
Marketing copy, emails, reports, titles, rewriting, and multilingual translation.
Understanding and extraction
Classification, entity extraction, sentiment, summaries, and structured information.
Coding tasks
Code explanation, generation, refactoring, testing, and error analysis.
Basic request
POST /v1/chat/completions
import os
from openai import OpenAI
client = OpenAI(
api_key=os.environ["APIQIK_API_KEY"],
base_url="<BASE_URL>/v1",
)
response = client.chat.completions.create(
model="<model-id>",
messages=[
{"role": "system", "content": "You are a professional product advisor."},
{"role": "user", "content": "Write three selling points for an API aggregation platform."},
],
temperature=0.6,
max_tokens=500,
)
print(response.choices[0].message.content)Model selection criteria
Models differ widely in context length, concurrency, and billing. Judge by these dimensions rather than guessing from the name:
| Dimension | What to confirm |
|---|---|
| Task type | Question answering, writing, summarization, translation, extraction, or code. |
| Context length | Whether the usable input plus output size covers your document length. |
| Capability tags | Streaming, tool calls, JSON output, multimodal input. |
| Latency and concurrency | Time to first token for interactive use; throughput for batch work. |
| Billing | Tiered input and output token prices. |
Model identifiers come from the model catalog; this page does not list a fixed set.
Query the current catalog
The catalog changes with channel availability, so trust the endpoint:
curl "${BASE_URL}/v1/public/models" \
-H "Authorization: Bearer ${APIQIK_API_KEY}"data[].id in the response is the exact string for the request's model field. The console model catalog shows capability tags, context length, and prices; check the catalog before choosing and never hard-code model names in clients.
Model capability card
Each language model in the catalog shows the following fields so developers do not choose by name alone.
| Field | Description |
|---|---|
| Model identifier | Exact, copyable string for the request's model field. |
| Context length | Combined input and output capacity. |
| Maximum output | Maximum generated tokens per request. |
| Capabilities | Streaming, tools, JSON, multimodal input, and other supported features. |
| Input/output prices | Separate token prices and billing definitions. |
| Suitable tasks | Recommended uses such as chat, code, translation, or long documents. |
| Limitations | Unsupported parameters, regions, and known differences. |
Selection guidance
- Prefer lower temperatures for factual answers and extraction; increase randomness where appropriate for creative writing.
- Compare quality, latency, and costs on real business samples, not one public benchmark score.
- Record model IDs and evaluation versions, and rerun regression tests after model updates.
- Set maximum output, timeouts, bounded retries, and traceable logging in production.