> ## Documentation Index > Fetch the complete documentation index at: https://docs.laozhang.ai/llms.txt > Use this file to discover all available pages before exploring further. # GPT-6 API launch: Astra, Sol, and Luna > GPT-6 Astra, Sol, and Luna are live on LaoZhang API. Get the model IDs, standard and long-context prices, how to choose, and API examples. LaoZhang API now serves OpenAI's GPT-6 Astra, Sol, and Luna. All three are billed per token in the `default` group, and you call them through Chat Completions or Responses by setting `model` to the model ID. * **Published**: September 24, 2026 * **Last verified**: September 24, 2026 * **Status**: Active. All three models are available in the `default` token group. ## Key facts | Item | Details | | ------------------------- | -------------------------------------------------------------------------------------------------------------------------- | | Model IDs | `gpt-6-astra`, `gpt-6-sol`, `gpt-6-luna` | | OpenAI release date | `gpt-6-astra`: September 3, 2026; `gpt-6-sol` and `gpt-6-luna`: September 22, 2026 | | Available on LaoZhang API | `gpt-6-astra` since September 10, 2026; `gpt-6-sol` and `gpt-6-luna` in the pricing configuration as of September 24, 2026 | | Endpoints | Base URL `https://api.laozhang.ai/v1`; Chat Completions `POST /v1/chat/completions`, Responses `POST /v1/responses` | | Token group and billing | `default` group, usage-based (per token) | | Pricing | Standard (up to 272K input tokens) and long-context tiers; see [Pricing](#pricing) | | Long-context tier | More than 272K input tokens in one request | | Context length (OpenAI) | 1,050,000-token context window, with up to 922,000 input tokens and 128,000 output tokens | ## Pricing These rates come from LaoZhang API's public pricing as of September 24, 2026, in US dollars per 1M tokens. They match the Standard rates on [OpenAI's pricing page](https://developers.openai.com/api/docs/pricing). **Standard requests** (up to 272K input tokens per request): | Model | Input | Output | Cache read | Cache write | | ------------- | ----: | -----: | ---------: | ----------: | | `gpt-6-astra` | \$10 | \$50 | \$1 | \$12.5 | | `gpt-6-sol` | \$2 | \$10 | \$0.2 | \$2.5 | | `gpt-6-luna` | \$0.1 | \$0.5 | \$0.01 | \$0.125 | **Long-context requests** (more than 272K input tokens per request): | Model | Input | Output | Cache read | Cache write | | ------------- | ----: | -----: | ---------: | ----------: | | `gpt-6-astra` | \$20 | \$75 | \$2 | \$25 | | `gpt-6-sol` | \$4 | \$15 | \$0.4 | \$5 | | `gpt-6-luna` | \$0.2 | \$0.75 | \$0.02 | \$0.25 | Once a request has more than 272K input tokens, the whole request is billed at the long-context rates. For current rates, see the [model and pricing catalog](/en/models) and the [console pricing page](https://api.laozhang.ai/account/pricing). ## Choose a model OpenAI suggests weighing the reasoning your task needs against latency and cost: * **`gpt-6-astra`**: OpenAI's most capable model, aimed at the hardest end-to-end work such as complex reasoning, coding, research, and document creation. It has the highest rates of the three. * **`gpt-6-sol`**: built for complex coding and agentic workflows, and suited to demanding tasks that need strong reasoning. * **`gpt-6-luna`**: the model OpenAI describes as its most efficient, for focused, repeatable tasks at high volume. It has the lowest rates of the three. Before you commit to one, run a sample of your own requests through the candidates and compare quality, latency, and token usage. [Choose a model](/en/api-reference/chat-completions#choose-a-model) describes how. ## Who is affected * **Affected**: projects that call text models through LaoZhang API's OpenAI-compatible endpoint. An API key in the `default` group with the Usage first or Usage-based billing mode can call GPT-6, so you can start testing by changing `model` to a GPT-6 ID. * **Not affected**: existing models. GPT-5.6, GPT-5.4, and other models remain available, and current integrations need no changes. ## Call GPT-6 Create an API key in [Token management](https://api.laozhang.ai/token) in the `default` group, with its billing mode set to Usage first (按量优先, recommended) or Usage-based (按量计费). Then export the key in your terminal: ```bash theme={null} export LAOZHANG_API_KEY="YOUR_LAOZHANG_API_KEY" ``` You can call GPT-6 through Chat Completions or Responses. These examples first send one message to `gpt-6-sol` with Chat Completions. For Python, run `python -m pip install openai` first. ```bash theme={null} curl https://api.laozhang.ai/v1/chat/completions \ -H "Authorization: Bearer $LAOZHANG_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "gpt-6-sol", "messages": [ {"role": "user", "content": "Explain what a database index does in three sentences."} ] }' ``` ```python theme={null} import os from openai import OpenAI client = OpenAI( api_key=os.environ["LAOZHANG_API_KEY"], base_url="https://api.laozhang.ai/v1", timeout=300.0, max_retries=0, ) response = client.chat.completions.create( model="gpt-6-sol", messages=[{"role": "user", "content": "Explain what a database index does in three sentences."}], ) print(response.choices[0].message.content) ``` The reply text is in `choices[0].message.content`, which the Python example prints. To use `gpt-6-astra` or `gpt-6-luna`, change only `model`. The Python example turns off automatic retries so that your application decides when to send a request again. [Call logs](https://api.laozhang.ai/log) show the tokens and the charge for each request. To use the Responses API, send `input` to `POST /v1/responses`: ```bash theme={null} curl https://api.laozhang.ai/v1/responses \ -H "Authorization: Bearer $LAOZHANG_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "gpt-6-sol", "input": "Explain what a database index is in three sentences." }' ``` ```python theme={null} import os from openai import OpenAI client = OpenAI( api_key=os.environ["LAOZHANG_API_KEY"], base_url="https://api.laozhang.ai/v1", timeout=300.0, max_retries=0, ) response = client.responses.create( model="gpt-6-sol", input="Explain what a database index is in three sentences.", ) print(response.output_text) ``` In the Python SDK, `response.output_text` joins the reply text. In the raw JSON, the text is in `content[].text` of the `output` item whose `type` is `message`. ## Check your request parameters before you switch All three GPT-6 models are reasoning models. OpenAI's [Using GPT-6](https://developers.openai.com/api/docs/guides/latest-model) guide lists the parameters to check when you move to GPT-6. These rules apply to Chat Completions requests: * **Reasoning effort** (`reasoning_effort`): * `gpt-6-sol` and `gpt-6-luna` accept `none`, `low`, `medium` (the default), `high`, `xhigh`, and `max`. * `gpt-6-astra` does not accept `none`; use `low` instead. * If your requests use `minimal`, switch to `low` for any of the three models and compare the results. * **Sampling parameters**: unless reasoning effort is `none`, remove `temperature`, `top_p`, `logprobs`, and `top_logprobs` from the request. Because `gpt-6-astra` has no `none` setting, never send these parameters to it. * **Tool calling**: * Tool calling with `gpt-6-astra` requires the Responses API. * `gpt-6-sol` and `gpt-6-luna` support function calling in Chat Completions only when `reasoning_effort` is `none`. * To use tools with `gpt-6-astra`, or reasoning together with function calling on Sol or Luna, send the request to LaoZhang API's Responses endpoint (`POST /v1/responses`). ## Frequently asked questions ### Can I call GPT-6 through the Responses API? Yes. LaoZhang API supports OpenAI's Responses API: set the base URL to `https://api.laozhang.ai/v1` and call `client.responses.create` as in OpenAI's GPT-6 guide; see the examples above. Tool calling with `gpt-6-astra` requires the Responses API, while plain conversations work with either endpoint. ### How are requests with more than 272K input tokens billed? The whole request is billed at the long-context rates, not just the tokens above 272K. Input, output, cache reads, and cache writes all use the long-context table above. Compared with the standard rates, input and cache rates are doubled and output costs 1.5 times as much, the same rule OpenAI applies. To estimate cost, use a request's full input token count to decide which tier applies. ### I already use GPT-5.6. Do I need to switch? No. `gpt-5.6-luna`, `gpt-5.6-terra`, and `gpt-5.6-sol` remain in the pricing configuration, and existing integrations keep working. If you plan to switch, compare both generations on representative requests for quality, latency, and cost, then check reasoning effort, sampling parameters, and tool calling as described above. The [model and pricing catalog](/en/models) lists current rates for both generations. ### What if the model doesn't exist or isn't allowed? Check the model ID, token group, and endpoint, in this order: 1. The model ID is copied exactly (`gpt-6-astra`, `gpt-6-sol`, or `gpt-6-luna`). 2. Your API key is in the `default` group with the Usage first or Usage-based billing mode, because a Per-call key can call only per-call models. 3. The request goes to `POST /v1/chat/completions` or `POST /v1/responses`. If it still fails, follow the steps in [model availability](/en/faq/model-availability). ### How do I check what my GPT-6 requests cost? Multiply a request's token counts by the current rates and compare the result with the actual charge: 1. Confirm the current rates on the [console pricing page](https://api.laozhang.ai/account/pricing). 2. Pick a recent request in [call logs](https://api.laozhang.ai/log) and work out its expected charge from its token counts. Use the long-context rates if the request had more than 272K input tokens. 3. Compare the result with the actual charge. If the amounts differ, email [hi@laozhang.ai](mailto:hi@laozhang.ai) with the request time and model ID, and do not include your API key. ## Sources and related links ### OpenAI * [OpenAI API changelog](https://developers.openai.com/api/docs/changelog): the September 3, 2026 entry for GPT-6 Astra and the September 22, 2026 entry for GPT-6 Sol and Luna * [OpenAI API pricing](https://developers.openai.com/api/docs/pricing): standard and long-context rates for all three models * Model pages for [GPT-6 Astra](https://developers.openai.com/api/docs/models/gpt-6-astra), [GPT-6 Sol](https://developers.openai.com/api/docs/models/gpt-6-sol), and [GPT-6 Luna](https://developers.openai.com/api/docs/models/gpt-6-luna): context window, maximum output, reasoning effort, and pricing rules * [Using GPT-6](https://developers.openai.com/api/docs/guides/latest-model): how the three models differ, and the parameter and tool-calling rules for switching ### LaoZhang API LaoZhang API rates, token groups, and endpoints come from the public pricing configuration as of September 24, 2026: * [Model and pricing catalog](/en/models): current rates, including long-context tiers * [Console pricing page](https://api.laozhang.ai/account/pricing): the rates that apply to your account * [Token management](https://api.laozhang.ai/token): create an API key and set its group and billing mode * [Call logs](https://api.laozhang.ai/log): token counts and the actual charge for each request ### Related documentation * [Use OpenAI models with LaoZhang API](/en/api-reference/openai) * [OpenAI protocol: Chat Completions](/en/api-reference/chat-completions) * [Text generation API: protocols and models](/en/api-capabilities/text-generation) * [GPT-5.6 Sol price cut](/en/announcements/gpt-5-6-sol-price-cut-2026-09) * [Latest announcements](/en/changelog)