This quick start shows how to make the first verified Ling 3.0 Tiny chat completions call on Novita AI. If you want the short version first: use https://api.novita.ai/openai as the base URL, inclusionai/ling-3.0-tiny as the model ID, and POST https://api.novita.ai/openai/v1/chat/completions as the request path.
When to Use This Quick Start
Use this page when your main question is practical setup: how to confirm the Ling 3.0 Tiny model ID, how to hit the OpenAI-compatible endpoint, and how to make the first successful chat completions request on Novita AI.
If your main question is model positioning, launch context, or whether this model is the right fit compared with another option, that belongs in a 4.2 or 4.3 page rather than this quick-start page.
Step 1: Get Your Novita API Key
Create a Novita AI API key and keep it out of source control. For a local smoke test, export it in your shell:
export NOVITA_API_KEY="your_api_key"In production, store the key in your secret manager instead of hardcoding it in application code.
Step 2: Confirm Model ID and Endpoint
As checked on August 10, 2026, the implementation details to verify before writing code are:
| Field | Value |
|---|---|
| Model ID | inclusionai/ling-3.0-tiny |
| Base URL | https://api.novita.ai/openai |
| Chat completions endpoint | POST https://api.novita.ai/openai/v1/chat/completions |
| Context window | 262,144 tokens |
| Max output tokens | 32,768 |
| Listed features | Function calling, prompt caching, Thinking mode, Instant mode |
| Pricing | Check the live model page and Novita pricing page before launch |
For this quick start, the important point is that Ling 3.0 Tiny uses the same OpenAI-compatible request pattern as other Novita-hosted chat models.
Step 3: Send Your First Request
Start with a small text-only request. This verifies authentication, model routing, and basic response parsing before you try long-context prompts or feature-specific fields.
curl "https://api.novita.ai/openai/v1/chat/completions" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer ${NOVITA_API_KEY}" \
-d '{
"model": "inclusionai/ling-3.0-tiny",
"messages": [
{
"role": "system",
"content": "You are a concise technical assistant."
},
{
"role": "user",
"content": "Write a 3-step checklist for testing a long-context chat app."
}
],
"max_tokens": 1024,
"temperature": 0.2
}'If this request returns a normal chat completion object, your first integration step is working.
Step 4: Read the Response
For the first smoke test, confirm these fields:
modelmatchesinclusionai/ling-3.0-tinychoices[0].message.contentcontains the answer textusageis present when returned for your request path- your client handles non-200 responses without exposing the API key
Do not move to a larger integration until your app can reliably extract the assistant message and log the basic request outcome.
Step 5: Check Pricing, Limits, and Common Errors
Before production rollout, re-check:
- live pricing on the model page and Novita pricing page
- context and output limits for your exact use case
- whether the features you need are available on the current model page
- whether your SDK wrapper preserves the exact base URL and path
Common early mistakes:
- wrong model ID
- wrong base URL
- putting the endpoint path into a client that already expects a base URL
- testing with a too-large prompt before the smoke test passes
Python Example
import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.novita.ai/openai",
api_key=os.environ["NOVITA_API_KEY"],
)
response = client.chat.completions.create(
model="inclusionai/ling-3.0-tiny",
messages=[
{"role": "system", "content": "You are a concise technical assistant."},
{"role": "user", "content": "Summarize the best first test for a new Novita API client."},
],
max_tokens=1024,
temperature=0.2,
)
print(response.choices[0].message.content)Use this version when your application already uses the OpenAI Python SDK pattern and you only need to swap in the Novita base URL and Ling model ID.
cURL Example
curl "https://api.novita.ai/openai/v1/chat/completions" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer ${NOVITA_API_KEY}" \
-d '{
"model": "inclusionai/ling-3.0-tiny",
"messages": [
{
"role": "system",
"content": "You are a concise technical assistant."
},
{
"role": "user",
"content": "Summarize the first checks for a new chat completions integration."
}
],
"max_tokens": 512,
"temperature": 0.2
}'This is the safest first request when you want to confirm that the route, key, and model ID are all accepted.
Key Parameters
For the first integration pass, these are the parameters to pay attention to:
model: use the exact Novita model IDmessages: send standard chat role/content objectsmax_tokens: keep this small for the smoke testtemperature: keep it low if you want stable test output
Once the first request works, you can move on to function calling, prompt caching, or mode-specific testing.
Troubleshooting
401 Unauthorized
Usually means the API key is missing, malformed, or not loaded into the current shell or runtime.
404 or model not found
Usually means the model string is wrong. Use the exact value inclusionai/ling-3.0-tiny.
Wrong endpoint behavior
If your SDK takes a base URL, set it to https://api.novita.ai/openai, not the full /v1/chat/completions path.
Large prompt failures
Do not start with a near-limit context test. Confirm the short request first, then scale prompt size step by step.
FAQ
What model ID does Ling 3.0 Tiny use on Novita AI?
Use inclusionai/ling-3.0-tiny.
What endpoint should I call for chat completions?
Use POST https://api.novita.ai/openai/v1/chat/completions.
What base URL should I use in OpenAI-compatible clients?
Use https://api.novita.ai/openai.
How much context does Ling 3.0 Tiny support on Novita AI?
Novita currently lists a 256K context window.
What is the listed max output length?
Novita currently lists 32,768 max output tokens.
Sources:
- Novita model page: https://novita.ai/models/model-detail/inclusionai-ling-3.0-tiny
- Novita docs index: https://novita.ai/docs/llms.txt
- Novita chat completions API: https://novita.ai/docs/api-reference/model-apis-llm-create-chat-completion
- Checked: August 10, 2026
