OpenAI GPT 6 Astra: AI 大语言模型
Reason through complex text, image, and file inputs with OpenAI's GPT 6 Astra. Adjustable reasoning depth and web search. No subscription.
Reason through complex text, image, and file inputs with OpenAI's GPT 6 Astra. Adjustable reasoning depth and web search. No subscription.
关于此模型
GPT 6 Astra is OpenAI's advanced multimodal reasoning model, built for tasks that mix text, image, and file inputs in a single conversation. It supports adjustable reasoning effort so you can trade latency for deeper multi-step logic, and can optionally call an integrated web search tool for real-time, up-to-date answers. Astra is well suited to complex analysis, coding, and research assistant workloads that need more careful reasoning than lighter chat models like GPT 5.6 Luna can reliably provide.
成本分析
| 提供商 | 费用 | 备注 |
|---|---|---|
| muapiapp | $10.00/M 输入, $50.00/M 输出 | Token-based billing, matching OpenAI's own published rate. Minimum $0.0008 per call. |
| Fal.ai | 暂不可用 | No dedicated GPT 6 Astra endpoint at this time. |
| Replicate | 暂不可用 | No dedicated GPT 6 Astra endpoint at this time. |
Token-based billing, matching OpenAI's own published rate. Minimum $0.0008 per call.
No dedicated GPT 6 Astra endpoint at this time.
No dedicated GPT 6 Astra endpoint at this time.
** 竞品价格根据相似模型架构和使用层级估算。
配置参数
| 参数 | 类型 | 描述 | 默认值 |
|---|---|---|---|
| 提示词 | string | 用户消息或指令。 | Summarize this article in three bullet points. |
| 图像 URL | string | 多模态请求的可选图像 URL。 | undefined |
| 系统提示词 | string | 可选的系统级指令，用于引导模型行为。 | You are an expert tutor. |
| 网页搜索开关 | boolean | 启用网页搜索工具以获取实时信息。 | false |
| 推理强度 | 枚举（4 个选项） | 推理深度级别：low、medium、high 或 xhigh。 | low |
用户消息或指令。
Summarize this article in three bullet points.多模态请求的可选图像 URL。
undefined可选的系统级指令，用于引导模型行为。
You are an expert tutor.启用网页搜索工具以获取实时信息。
false推理深度级别：low、medium、high 或 xhigh。
low开发者文档
POST /api/v1/gpt-6-astra — returns a request_id; poll for the result via /api/v1/predictions/{id}/result.
POST /api/v1/gpt-6-astra/stream — returns a live Server-Sent Events stream (Content-Type: text/event-stream). Each chunk has the format data: {"choices":[{"delta":{"content":"text"}}]}, ending with data: [DONE].
curl -X POST "https://api.muapi.ai/api/v1/gpt-6-astra" \
-H "Authorization: Bearer $YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"prompt": "Walk through the tradeoffs of eventual consistency in a distributed database.",
"reasoning_effort": "high"
}'
常见问答
Astra is OpenAI's advanced reasoning model, built for complex analysis, coding, and research tasks that need careful multi-step logic rather than quick, shallow responses.
Pricing is token-based at $10.00 per million input tokens and $50.00 per million output tokens, with a minimum charge of $0.0008 per call.
Yes, Astra accepts mixed text and image (image_url) inputs in a single message, alongside an optional system_prompt.
Pass a reasoning_effort value of 'low', 'medium', 'high', or 'xhigh' to control how much internal reasoning the model performs before responding. Higher levels improve accuracy on hard problems but increase latency and output token usage.
Yes, set web_search_switch to true to let the model call an integrated web search tool for real-time information. Web search and custom function calling are mutually exclusive per request.
Use the streaming endpoint (/gpt-6-astra/stream) for interactive chat interfaces where the user expects real-time token-by-token feedback. Use the async endpoint (/gpt-6-astra) for background tasks, workflows, and automated scripts.