Apodex 1.1 Mini on Novita AI: API Quickstart and Pricing

Apodex 1.1 Mini is available on Novita AI through an OpenAI-compatible API. Set the model to apodex/apodex-1.1-mini, send requests to https://api.novita.ai/v3/openai, and use the model’s 256K-token context for long coding, tool-use, and agent workflows. As of October 1, 2026, Novita lists input and output pricing at $0 per million tokens as a launch promotion, so check the live model page before putting the model into a production budget.

What is Apodex 1.1 Mini?

Apodex 1.1 Mini is a 35B-parameter mixture-of-experts model in the Apodex 1.1 series. Novita describes it as using a Qwen3.5-MoE architecture and serving it on dedicated H200 GPU fleets with TP4 tensor parallelism, prefix-cache-aware routing, and MTP speculative decoding. Those deployment details are useful context for developers evaluating latency and throughput, but they are not a replacement for measuring your own prompts and tool workflows.

The model is a text-in, text-out chat model. It is available through Novita’s serverless API and supports the chat/completions, Anthropic, and Responses endpoints listed in the model catalog. For the shortest migration path, use the OpenAI-compatible chat/completions endpoint shown below.

The Apodex 1.1 Mini model page is the source of truth for the current model status, supported features, and playground access.

Apodex 1.1 Mini specifications

Specification	Value
Model ID	`apodex/apodex-1.1-mini`
Model family	Apodex 1.1
Architecture	Qwen3.5-MoE
Parameter count	35B
Context window	262,144 tokens (256K)
Maximum output	262,144 tokens (256K)
Input and output	Text in, text out
Features	Function calling, structured outputs, reasoning
API endpoints	Chat Completions, Anthropic, Responses
Hosting	Novita serverless API

Apodex 1.1 Mini’s 256K context is the most practical reason to test it with large repositories, long specifications, or multi-step agent traces. A large context does not mean every request should include the entire workspace: trim irrelevant files, keep tool results bounded, and measure both quality and latency on representative inputs.

How to call Apodex 1.1 Mini

You need a Novita API key. Create or manage one from Key Management, then keep it in an environment variable rather than putting it in source control.

Install the OpenAI Python client

pip install openai

Send a first request

The endpoint is OpenAI-compatible, so the standard OpenAI client only needs a Novita base URL, API key, and model ID.

import os

from openai import OpenAI


client = OpenAI(
    api_key=os.environ["NOVITA_API_KEY"],
    base_url="https://api.novita.ai/v3/openai",
)

response = client.chat.completions.create(
    model="apodex/apodex-1.1-mini",
    messages=[
        {
            "role": "system",
            "content": "You are a concise coding assistant.",
        },
        {
            "role": "user",
            "content": "Explain why a database index can speed up a query.",
        },
    ],
    max_tokens=512,
)

print(response.choices[0].message.content)

Set the key before running the script:

export NOVITA_API_KEY="your_novita_api_key"
python apodex_example.py

For a raw HTTP integration, the equivalent request is:

curl https://api.novita.ai/v3/openai/chat/completions \
  -H "Authorization: Bearer $NOVITA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "apodex/apodex-1.1-mini",
    "messages": [
      {"role": "user", "content": "Give me three names for a log parser."}
    ],
    "max_tokens": 128
  }'

See the LLM API documentation for the broader request format and supported client patterns.

Streaming responses

For chat interfaces, set stream=True and print each delta as it arrives. Streaming changes how your application receives the response; it does not change the model ID or authentication setup.

stream = client.chat.completions.create(
    model="apodex/apodex-1.1-mini",
    messages=[{"role": "user", "content": "Write a short commit message for a bug fix."}],
    max_tokens=128,
    stream=True,
)

for chunk in stream:
    text = chunk.choices[0].delta.content
    if text:
        print(text, end="", flush=True)

How to use tool calling

The model catalog lists native function calling and structured outputs as supported features. A tool definition tells the model what your application can do; your application still owns execution, validation, permissions, and the final tool result.

tools = [
    {
        "type": "function",
        "function": {
            "name": "get_weather",
            "description": "Get the current weather for a city.",
            "parameters": {
                "type": "object",
                "properties": {
                    "city": {"type": "string"},
                },
                "required": ["city"],
                "additionalProperties": False,
            },
        },
    }
]

response = client.chat.completions.create(
    model="apodex/apodex-1.1-mini",
    messages=[
        {"role": "user", "content": "What's the weather in Seattle?"},
    ],
    tools=tools,
    tool_choice="auto",
)

message = response.choices[0].message
if message.tool_calls:
    tool_call = message.tool_calls[0]
    print(tool_call.function.name)
    print(tool_call.function.arguments)

Treat the arguments as untrusted model output. Parse the JSON, validate the city against your own input rules, enforce authorization before calling an external service, and send the tool result back in a follow-up request. Do not let a model-generated argument directly become a shell command, SQL statement, or privileged API call.

The model page also lists reasoning support. The exact reasoning controls and response behavior can vary by client and endpoint, so start with the default behavior and confirm the current API documentation before depending on a provider-specific reasoning parameter.

What does Apodex 1.1 Mini cost?

As of October 1, 2026, Novita’s model catalog and pricing page show Apodex 1.1 Mini as free for both input and output tokens:

Usage	Current listed price
Input	$0 per million tokens
Output	$0 per million tokens

This is a current launch promotion, not a guarantee of permanent free access. Pricing, rate limits, and availability can change. Check the live Novita pricing page and the Apodex model page immediately before estimating production costs or committing to a long-running workload.

Even at a $0 listed price, set application-level budgets and rate limits. Large-context requests can consume substantial compute and may be subject to account, endpoint, or promotional limits. Record token usage from API responses when available, and keep a fallback model configured if your application depends on continuous access.

Who should use it?

Apodex 1.1 Mini is a good candidate for:

Coding assistants that need to inspect long files or repository context.
Agent workflows that call tools and return structured data.
Prototypes where a 256K context window is useful but self-hosting a large model is unnecessary.
Evaluations of MoE inference, long-context behavior, and reasoning workflows on a hosted API.

It may not be the right fit when you need image or audio input, a fixed long-term price contract, or a model-specific feature that Novita does not list for this endpoint. The catalog currently lists text-only input and output, so multimodal requests should use a model whose page explicitly supports those modalities.

For a production evaluation, compare answer quality on your real prompts, tool-call validity, time to first token, total latency, and context trimming behavior. A single short prompt cannot tell you whether a long-context model is a good fit for your application.

Frequently asked questions

What is the Apodex 1.1 Mini model ID?

Use apodex/apodex-1.1-mini in the request body.

How much context does Apodex 1.1 Mini support?

Novita lists a 262,144-token context window, displayed as 256K on the model and pricing pages. The listed maximum output is also 262,144 tokens.

Is Apodex 1.1 Mini free on Novita AI?

The input and output prices are currently listed as $0 per million tokens as of October 1, 2026. Treat that as a launch promotion and verify the live pricing page before publishing a cost estimate or deploying at scale.

Does it support function calling?

Yes. The model catalog lists function calling and structured outputs as supported features. Your application must still validate tool arguments and execute tools securely.

Which API should I use first?

Use the OpenAI-compatible Chat Completions endpoint at https://api.novita.ai/v3/openai/chat/completions for the simplest integration. The model catalog also lists Anthropic and Responses endpoints.

Where can I try Apodex 1.1 Mini?

Open the Apodex 1.1 Mini model page to check the current status and launch the playground.