> ## Documentation Index
> Fetch the complete documentation index at: https://apixo.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# OpenAI Responses API Gateway: Models, Pricing & Code

> Integrate ten GPT models through APIXO's OpenAI Responses API gateway. Compare pricing and use copy-ready cURL, JavaScript, Python, or MCP instructions.

Use one OpenAI Responses-compatible integration for all ten GPT models available through APIXO. Configure the APIXO base URL and API key once, then switch models by changing only the `model` value.

<Info>
  This page includes the complete APIXO model and pricing snapshot verified on August 3, 2026. Model availability and pricing can change, so check the [APIXO OpenAI model page](https://apixo.ai/models/llm/openai) before making billing-sensitive decisions.
</Info>

<Tip>
  **Choose your integration path:** [Use the Quickstart](#quickstart) to integrate manually, or [let an AI implement it](#let-an-ai-implement-it) by copying this page or using APIXO MCP.
</Tip>

## Compare GPT models and pricing

Choose the model that fits the workload, then use its exact model ID in any example on this page.

| Model         | Model ID        | Recommended for                                                       |  Input | Output | Cache write | Cache read | Savings |
| ------------- | --------------- | --------------------------------------------------------------------- | -----: | -----: | ----------: | ---------: | ------: |
| GPT-5.6 Sol   | `gpt-5.6-sol`   | Most demanding GPT-5.6 coding and professional workloads              |    \$3 |   \$18 |      \$3.75 |      \$0.3 |     40% |
| GPT-5.6 Terra | `gpt-5.6-terra` | Balanced GPT-5.6 capability and cost for everyday professional work   |  \$1.5 |    \$9 |     \$1.875 |     \$0.15 |     40% |
| GPT-5.6 Luna  | `gpt-5.6-luna`  | Cost-efficient GPT-5.6 option for routine and higher-volume workloads |  \$0.6 |  \$3.6 |      \$0.75 |     \$0.06 |     40% |
| GPT-5.5       | `gpt-5.5`       | Most complex coding and professional work                             |    \$3 |   \$18 |       \$0.3 |          — |     40% |
| GPT-5.4       | `gpt-5.4`       | Complex professional work at a lower cost than GPT-5.5                |  \$1.5 |    \$9 |      \$0.15 |          — |     40% |
| GPT-5.4 Mini  | `gpt-5.4-mini`  | High-volume coding, computer use, and subagent workloads              | \$0.45 |  \$2.7 |     \$0.045 |          — |     40% |
| GPT-5.3-Codex | `gpt-5.3-codex` | Agentic coding in Codex or similar environments                       | \$1.05 |  \$8.4 |     \$0.105 |          — |     40% |
| GPT-5.2       | `gpt-5.2`       | Previous-generation complex professional workloads                    | \$1.05 |  \$8.4 |     \$0.105 |          — |     40% |
| GPT-5.2-Codex | `gpt-5.2-codex` | Long-horizon agentic coding on the GPT-5.2 family                     | \$1.05 |  \$8.4 |     \$0.105 |          — |     40% |
| GPT-5 Mini    | `gpt-5-mini`    | Well-defined, cost-sensitive, low-latency, high-volume tasks          | \$0.15 |  \$1.2 |     \$0.015 |          — |     40% |

Prices are in USD per 1 million tokens. `—` means APIXO does not publish a separate cache-read price for that model. Prices verified: 2026-08-03.

<Tip>
  Start with GPT-5.6 Sol for the most demanding GPT-5.6 workloads, GPT-5.6 Terra for a balanced default, or GPT-5.6 Luna when GPT-5.6 access at a lower price matters more. The earlier GPT-5.x and Codex models remain available for existing integrations and model-specific workloads; GPT-5 Mini is still the lowest-cost option in this table.
</Tip>

## What to change

| Setting        | APIXO value                                                                |
| -------------- | -------------------------------------------------------------------------- |
| SDK Base URL   | `https://llm.apixo.ai/v1`                                                  |
| Raw HTTP path  | `POST /v1/responses`                                                       |
| Request URL    | `https://llm.apixo.ai/v1/responses`                                        |
| API key        | `APIXO_API_KEY` from [APIXO API Keys](https://apixo.ai/dashboard/api-keys) |
| API key header | `Authorization`                                                            |
| API key value  | `Bearer $APIXO_API_KEY`                                                    |
| Model location | Request body `model`                                                       |

## Quickstart

```bash theme={null}
export APIXO_API_KEY="your_apixo_api_key"
```

<Tabs>
  <Tab title="cURL">
    ```bash theme={null}
    curl -X POST "https://llm.apixo.ai/v1/responses" \
      -H "Authorization: Bearer $APIXO_API_KEY" \
      -H "Content-Type: application/json" \
      -d '{
        "model": "gpt-5.6-sol",
        "input": "Say hello in one sentence."
      }'
    ```
  </Tab>

  <Tab title="JavaScript">
    ```javascript theme={null}
    import OpenAI from "openai";

    const client = new OpenAI({
      apiKey: process.env.APIXO_API_KEY,
      baseURL: "https://llm.apixo.ai/v1"
    });

    const response = await client.responses.create({
      model: "gpt-5.6-sol",
      input: "Say hello in one sentence."
    });

    console.log(response.output_text);
    ```
  </Tab>

  <Tab title="Python">
    ```python theme={null}
    import os
    from openai import OpenAI

    client = OpenAI(
        api_key=os.environ["APIXO_API_KEY"],
        base_url="https://llm.apixo.ai/v1",
    )

    response = client.responses.create(
        model="gpt-5.6-sol",
        input="Say hello in one sentence.",
    )

    print(response.output_text)
    ```
  </Tab>
</Tabs>

<Tip>
  The examples use `gpt-5.6-sol`. To use any other model in the comparison table, change only the `model` value.
</Tip>

## Let an AI implement it

### Copy this page to an AI

Use the **Copy page** button in the page menu, paste the copied documentation into your AI coding tool, and add this prompt:

```text theme={null}
Integrate the APIXO OpenAI Responses API into my existing project using the
documentation below.

Requirements:
- Use model gpt-5.6-sol unless my project already specifies another supported model.
- Read APIXO_API_KEY from an environment variable and keep it server-side.
- Preserve my current language, framework, and project structure.
- Add the smallest working integration and one minimal verification request.
- Do not invent endpoints, model IDs, or request fields that are not documented.
```

Replace `gpt-5.6-sol` in the prompt with any model ID from the comparison table.

### Use APIXO MCP

[APIXO MCP](/docs/integrations/mcp) reads APIXO's published machine-readable model catalog and schemas; it does not scrape this webpage in real time. The `openai-responses` schema includes the gateway endpoint, supported model IDs, model guidance, and the pricing snapshot shown above, so an MCP-connected AI can discover the current integration details and write the code for you.

After [installing APIXO MCP](/docs/integrations/mcp/installation), ask your AI:

```text theme={null}
Use the APIXO MCP tools to find the openai-responses model schema. Compare the
supported GPT models and prices, choose the best model for my stated workload,
then integrate the APIXO Responses API into this project. Keep the API key
server-side and add a minimal verification request.
```

APIXO MCP helps your AI write the integration; it does not replace the API connection used by your application.

## Streaming

```bash theme={null}
curl -N -X POST "https://llm.apixo.ai/v1/responses" \
  -H "Authorization: Bearer $APIXO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.6-sol",
    "stream": true,
    "input": "Write a short changelog summary."
  }'
```

## Supported OpenAI models

Use one of these exact ids in the request body `model` field.

* `gpt-5.6-sol`
* `gpt-5.6-terra`
* `gpt-5.6-luna`
* `gpt-5.5`
* `gpt-5.4`
* `gpt-5.4-mini`
* `gpt-5.3-codex`
* `gpt-5.2`
* `gpt-5.2-codex`
* `gpt-5-mini`

To switch models, keep the same endpoint, API key, and request format, then change only `model`.

## Troubleshooting

* `401`: missing or invalid API key
* `400`: invalid Responses API request body, misspelled model id, or unsupported model
* `429`: rate limit or temporary capacity limit
* `502` / `504`: upstream provider error or timeout

Retry `429`, `502`, and `504` with backoff. Fix `400` and `401` before retrying.

## Related links

* [Chat API Overview](/docs/llm)
* [Supported Models](/docs/llm/supported-models)
* [LLM Authentication](/docs/llm/authentication)
* [LLM Streaming](/docs/llm/streaming)
* [LLM Errors and Retry](/docs/llm/errors-retry)
* [Install APIXO MCP](/docs/integrations/mcp/installation)
* [Explore OpenAI models on APIXO](https://apixo.ai/models/llm/openai)
* [Pricing](https://apixo.ai/pricing)

## Open a model on APIXO

These links open the selected model on APIXO's OpenAI gateway page, where you can review its current pricing and copy model-specific code examples.

* [GPT-5.6 Sol](https://apixo.ai/models/llm/openai?model=gpt-5.6-sol)
* [GPT-5.6 Terra](https://apixo.ai/models/llm/openai?model=gpt-5.6-terra)
* [GPT-5.6 Luna](https://apixo.ai/models/llm/openai?model=gpt-5.6-luna)
* [GPT-5.5](https://apixo.ai/models/llm/openai?model=gpt-5.5)
* [GPT-5.4](https://apixo.ai/models/llm/openai?model=gpt-5.4)
* [GPT-5.4 Mini](https://apixo.ai/models/llm/openai?model=gpt-5.4-mini)
* [GPT-5.3-Codex](https://apixo.ai/models/llm/openai?model=gpt-5.3-codex)
* [GPT-5.2](https://apixo.ai/models/llm/openai?model=gpt-5.2)
* [GPT-5.2-Codex](https://apixo.ai/models/llm/openai?model=gpt-5.2-codex)
* [GPT-5 Mini](https://apixo.ai/models/llm/openai?model=gpt-5-mini)