模型与能力Models & Capabilities
平台聚合国内外主流模型厂商。你可以先按能力选择接口,再查询当前密钥可调用模型,快速确定合适的 model 与调用路径。The platform aggregates leading model providers from home and abroad. Pick an endpoint by capability first, then query the models available to your current key to quickly determine the right model and calling path.
按能力选择入口Entry Points by Capability
| 能力Capability | 推荐接口Recommended Endpoint | 适合场景Best For |
|---|---|---|
| 文本生成:统一聊天Text Generation: Unified Chat | POST /v1/chat/completions | 默认推荐入口,适合统一调用 GPT、Claude、Gemini 等主流聊天模型The default recommended entry point for uniformly calling mainstream chat models such as GPT, Claude, and Gemini |
| 图片生成Image Generation | POST /v1/images/generations | 文生图 / 图生图Text-to-image / image-to-image |
| 语音合成Speech Synthesis | POST /v1/audio/speech | 文本转语音,按字符计费Text-to-speech, billed by character |
| 语音转写Speech Transcription | POST /v1/audio/transcriptions | 音频文件识别,按秒计费Audio file recognition, billed by second |
| 视频生成Video Generation | POST /v1/videos/generations | 异步任务,提交后轮询查询,支持文生视频 / 图生视频Async task: submit and poll for results; supports text-to-video / image-to-video |
| 模型查询Models | GET /v1/models | 查询当前密钥可调用的模型Query models callable with the current key |
模型查询接口Models Endpoint
用于查询当前密钥可调用模型:Used to query the models available to the current key:
| 接口Endpoint | 用途Purpose | 适合场景Best For |
|---|---|---|
GET /v1/models | 查询当前密钥可访问的模型列表List the models accessible with the current key | 接入前确认某个密钥当前能用哪些模型Confirm which models a key can use before integration |
models.shcURL
curl {BASE_URL}/v1/models \
-H "Authorization: Bearer YOUR_API_KEY"
模型选择建议Model Selection Tips
- 最快接入聊天能力Fastest way to add chat capability — 先用
/v1/chat/completions,优先从通用模型开始。 — Start with/v1/chat/completionsand begin with general-purpose models. - 需要生成图片Image generation — 使用
/v1/images/generations,按生成张数计费。 — Use/v1/images/generations, billed per generated image. - 需要生成视频Video generation — 使用
/v1/videos/generations,按实际生成秒数与分辨率档位计费。 — Use/v1/videos/generations, billed by actual generated seconds and resolution tier. - 先查目录再决定具体模型Check the catalog before choosing a model — 当你不确定权限或模型开放情况时,先查
GET /v1/models,再确定业务侧实际使用的model值。 — When unsure about permissions or model availability, queryGET /v1/modelsfirst, then decide themodelvalue your application actually uses.
使用说明Usage Notes
- 想先快速接入:优先使用
POST /v1/chat/completions。To integrate quickly: start withPOST /v1/chat/completions. - 想先确认某个密钥能调用哪些模型:使用
GET /v1/models。To confirm which models a key can call: useGET /v1/models. - 想生成视频:使用
POST /v1/videos/generations,提交后记录任务 ID 并轮询查询。To generate a video: usePOST /v1/videos/generations, save the task ID after submission, and poll for status. - 实际可用模型以你的账户权限和当前开放列表为准,可在「模型广场」查看。The models actually available depend on your account permissions and the current public catalog; see the Model Square.
切换供应商模型只需更改
model 参数,请求结构保持不变;不同模型的价格与计费方式可能不同,见「计费与配额」模块。Switching provider models only requires changing the model parameter; the request structure stays the same. Pricing and billing methods may differ across models — see the Billing & Quota module.下一步Next Steps
- 快速开始Quick Start — 打通第一条请求。 — Complete your first request.
- API 参考API Reference — 了解不同协议的接入选择与请求参数。 — Learn about protocol options and request parameters.
- 视频生成Video Generation — 查看视频模型选择与提示词建议。 — See video model selection and prompt tips.