diff --git a/docs/llmservice/auto-mode.md b/docs/llmservice/auto-mode.md new file mode 100644 index 0000000..a39fff9 --- /dev/null +++ b/docs/llmservice/auto-mode.md @@ -0,0 +1,93 @@ +# Auto Mode: Start Without Choosing a Model + +When working with different AI models, you may wonder which one is best for writing, coding, or a simple question. Auto Mode is designed to reduce that choice: describe what you want to accomplish, and B.AI selects an appropriate available model for the request. + +## What Is Auto Mode? + +Auto is an intelligent option in the B.AI Chat model selector. + +After selecting Auto, you can send messages, ask follow-up questions, and use conversation context as usual. The difference is that you do not need to select a specific model before each request. B.AI chooses from the models available to your account for each request. + +Each request is handled by one model and returns one complete response. The chat interface displays Auto rather than the routing process or selection rules. You can review the actual model, token usage, and cost in the request usage details and on the **Usage** page. + +:::info Choosing for the task + +Auto Mode aims to select a suitable model for the current task. It does not always select the largest model, the fastest model, or the lowest-priced model. + +::: + +## When to Use Auto Mode + +Auto Mode is a good choice when: + +* You are not sure which model to select and care more about the result than model comparison. +* Your tasks vary, such as researching, writing, and analyzing code in the same conversation. +* You want to start quickly with fewer setup steps. +* You are handling everyday tasks such as summarizing, polishing text, explaining concepts, organizing ideas, or drafting content. + +If you need to evaluate a particular model, reproduce results, or follow a workflow that requires a fixed model, select that model manually instead. + +## How to Use Auto Mode + +1. Open a new chat or enter an existing chat. +2. Click the model selector. +3. Select **Auto** from the model list. +4. Enter your question or task, then send it. +5. Continue the conversation as needed. + +Auto remains selected for later messages, but each request can independently use a different actual model. You do not need to ask B.AI to choose a model in your prompt. Instead, clearly describe your goal, context, constraints, and expected output. + +For example, instead of writing "Write an introduction," you could write: + +> Write a Chinese product introduction of no more than 150 characters for an AI image-organizing tool for designers. Keep the tone professional and concise, without exaggerated claims. + +## Auto Mode and Manual Model Selection + +| Item | Auto Mode | Manual Model Selection | +| :--- | :--- | :--- | +| Model selection | B.AI selects a model for each request | You select a specific model | +| Best for | Everyday tasks, varying work, or uncertain model choice | Model evaluation, reproducible results, or fixed workflows | +| Chat display | Displays Auto | Displays the selected model | +| Actual usage | Available in request usage details and on the Usage page | Available in request usage details and on the Usage page | +| Billing | Based on the actual model and usage | Based on the selected model and usage | + +If you do not have a specific model preference, start with Auto Mode. Switch to a manually selected model when your task requires a specific one. + +## Reviewing the Actual Model, Tokens, and Cost + +Auto does not show the actual model in the chat message itself, but every request has an execution record. Review the request usage details in the chat or open the **Usage** page to see: + +* The actual model used +* Input and output tokens +* Tools used +* The actual cost of the request + +Auto Mode does not add a separate model-selection charge. The final cost is determined by the model actually used, token usage, and any tool usage. Different requests may use different models and therefore have different costs. Refer to the usage record for the final amount. + +## Search and Tools + +When Web Search or read-only tools are enabled in a chat, Auto Mode can use the available tools when they are needed for the task. Availability depends on the current feature configuration and the capabilities of the model selected for the request. + +Tool calls and their related usage are also recorded on the **Usage** page. + +## Frequently Asked Questions + +### Will the same question always use the same model? + +Not necessarily. Available models, task context, and service configuration can change, so similar questions at different times may use different models. + +### Does Auto Mode call multiple models at the same time? + +No. Each Auto request selects one model and returns one complete response. + +### Can I see which model Auto Mode selected? + +Yes. The chat interface displays Auto to keep the conversation simple, while the actual model is available in the request usage details and on the **Usage** page. + +### Is Auto Mode always better than selecting a model manually? + +No. Auto Mode is useful when you do not have a specific model preference and want to reduce selection effort. Selecting a model manually gives you more control when you already know which model you need. + +### Is Auto Mode more expensive? + +Auto Mode has no additional charge. However, model pricing and token usage can differ by request. Review the actual cost in the **Usage** page. diff --git a/docs/llmservice/pricing-and-usage.md b/docs/llmservice/pricing-and-usage.md index 635df72..bd1806a 100644 --- a/docs/llmservice/pricing-and-usage.md +++ b/docs/llmservice/pricing-and-usage.md @@ -62,6 +62,10 @@ The main pricing table shows the currently effective standard reference price fo Prices shown in the documentation are B.AI standard reference prices for base billing purposes. B.AI may provide lower actual usage costs through top-up bonuses and account benefits. Specific prices, bonus Credits, and account benefits are subject to the platform display and final billing records. ::: +:::info Auto Mode billing +Auto Mode is not a separately billable model. Each request is billed based on the model actually used, token usage, and any tool usage. Review the request usage details or the **Usage** page for the final cost. See [Auto Mode](./auto-mode.md) for details. +::: + ### Cache Pricing Notes - **Cache Write:** The cost when tokens are first written into the prompt cache. Most providers charge no premium and use the same rate as standard input pricing. Claude models and GPT-5.6 models apply a 25% premium on cache writes; GPT-5.6 long-context cache write rates follow the corresponding long-context input tier. diff --git a/docs/llmservice/quick-start.md b/docs/llmservice/quick-start.md index 2a5d247..9c18087 100644 --- a/docs/llmservice/quick-start.md +++ b/docs/llmservice/quick-start.md @@ -18,7 +18,7 @@ Upon successful login, your wallet address will be displayed in the top right co ## 2. Start Your First Conversation Once inside the platform, you can immediately start interacting with the AI through our LLM Service. -* **Select AI Model:** In the chat interface, you will see the current default AI model. Click the model name to expand the list and select the model you wish to use. +* **Select AI Model:** In the chat interface, you will see the current default AI model. Click the model name to expand the list and select the model you wish to use. If you are unsure which model to choose, select **Auto** to let B.AI choose an appropriate available model for each request. See [Auto Mode](./auto-mode.md) for details. * **Send Message:** In the input box at the bottom of the page, enter your prompt or task, then click the send button or press Enter. * **Contextual Interaction:** The AI's response will be presented in a dialogue format. You can engage in multi-turn conversations, and the AI will respond based on the context. diff --git a/i18n/zh-Hans/docusaurus-plugin-content-docs/current.json b/i18n/zh-Hans/docusaurus-plugin-content-docs/current.json index 2b25739..4a44adb 100644 --- a/i18n/zh-Hans/docusaurus-plugin-content-docs/current.json +++ b/i18n/zh-Hans/docusaurus-plugin-content-docs/current.json @@ -63,6 +63,10 @@ "message": "快速入门", "description": "The label for the doc item Quick Start in sidebar docsSidebar" }, + "sidebar.docsSidebar.doc.Auto Mode": { + "message": "Auto 模式", + "description": "The label for the doc item Auto Mode in sidebar docsSidebar" + }, "sidebar.docsSidebar.doc.Pricing and Usage": { "message": "定价与用量", "description": "The label for the doc item Pricing and Usage in sidebar docsSidebar" diff --git a/i18n/zh-Hans/docusaurus-plugin-content-docs/current/llmservice/auto-mode.md b/i18n/zh-Hans/docusaurus-plugin-content-docs/current/llmservice/auto-mode.md new file mode 100644 index 0000000..be637f3 --- /dev/null +++ b/i18n/zh-Hans/docusaurus-plugin-content-docs/current/llmservice/auto-mode.md @@ -0,0 +1,95 @@ +# Auto 模式：不必先选模型，直接开始提问 + +面对不同的 AI 模型，你可能经常遇到这些问题：写作应该选哪个？代码任务换一个模型会不会更好？只是问一个简单问题，是否有必要反复比较？ + +Auto 模式就是为减少这些选择而设计的。你只需要描述想完成的任务，B.AI 会根据问题和当前可用能力，为本次请求选择合适的模型并生成答案。 + +## Auto 模式是什么 + +Auto 是 B.AI Chat 模型列表中的一个智能选择项。 + +选择 Auto 后，你仍然可以像平常一样发送消息、继续追问和使用对话上下文。不同之处在于，你不需要在提问前指定具体模型。B.AI 会在当前账户可使用的模型范围内，为每次请求完成模型选择。 + +每次请求仍由一个模型生成一个完整答案。聊天页面会显示 Auto，不展示内部的判断过程或选择规则；实际使用的模型、Token 和费用可以在对话中的用量详情和 **用量信息（Usage）** 页面查看。 + +:::info 为任务匹配模型 + +Auto 的目标是为当前任务选择合适的模型，而不是固定选择参数最多、速度最快或价格最低的模型。 + +::: + +## 哪些时候适合使用 Auto + +以下情况适合优先选择 Auto： + +* **不确定该选哪个模型：** 你更关心任务结果，不想先了解每个模型的差异。 +* **任务类型经常变化：** 同一段对话中可能先查资料，再写文案，之后又分析代码。 +* **希望快速开始：** 你想减少设置步骤，输入问题后直接获得回答。 +* **日常通用任务：** 例如总结内容、润色文字、解释概念、整理思路或生成初稿。 + +如果你需要测试某个模型的能力、复现实验结果，或者工作流程明确要求使用固定模型，更适合在模型列表中直接选择该模型。 + +## 如何使用 Auto + +1. 打开一个新会话或进入现有会话。 +2. 点击模型选择器。 +3. 在模型列表中选择 **Auto**。 +4. 输入你的问题或任务要求，然后发送消息。 +5. 根据回答继续追问。 + +后续消息会保持 Auto 为选中状态，但每次请求都可能独立选择不同的实际执行模型。你不需要在提示词中写“请帮我选择模型”。清楚描述目标、背景、限制条件和期望输出即可。 + +例如，与其只输入“写一段介绍”，可以写： + +> 为一款面向设计师的 AI 图片整理工具写一段 150 字以内的中文产品介绍，语气专业、简洁，不使用夸张承诺。 + +## Auto 与手动指定模型有什么区别 + +| 对比项 | Auto | 手动指定模型 | +| :--- | :--- | :--- | +| 模型选择 | B.AI 根据每次任务自动选择 | 由你选择指定模型 | +| 适合场景 | 日常使用、任务变化、不确定选型 | 模型测试、结果复现、固定工作流 | +| 聊天页展示 | 显示 Auto | 显示所选模型 | +| 实际用量 | 在对话用量详情与用量信息页面中查看 | 在对话用量详情与用量信息页面中查看 | +| 计费方式 | 按实际执行的模型和用量计费 | 按指定模型和实际用量计费 | + +如果你没有明确的模型偏好，可以从 Auto 开始；当任务对模型有明确要求时，再切换为手动指定。 + +## 查看实际模型、Token 和费用 + +Auto 不会在聊天正文中展示具体模型，但每次请求的执行记录仍然可查。你可以在对话中的用量详情或 **用量信息（Usage）** 页面查看： + +* 实际使用的模型 +* 输入和输出 Token +* 使用的工具 +* 本次请求产生的实际费用 + +Auto 不会额外收取一笔“选择模型”的费用。最终费用由实际执行的模型、Token 用量和工具使用情况决定。不同请求可能选择不同模型，因此费用也可能不同，请以用量记录为准。 + +## 搜索和工具 + +如果当前会话已经启用 Web Search 或只读工具，Auto 可以在任务需要时使用可用的工具能力。具体是否可用，取决于当前功能配置以及本次请求实际执行模型的能力。 + +工具调用及其相关用量同样会记录在 **用量信息（Usage）** 页面中。 + +## 常见问题 + +### 同一个问题每次都会选择相同的模型吗？ + +不一定。可用模型、任务上下文和服务配置可能发生变化，因此相似问题在不同时间可能由不同模型处理。 + +### Auto 会同时调用多个模型吗？ + +不会。每次 Auto 请求会选择一个模型，并返回一个完整答案。 + +### 可以看到 Auto 实际选择了哪个模型吗？ + +可以。聊天页面保持简洁，只显示 Auto；实际模型可在对话用量详情与 **用量信息（Usage）** 页面的请求记录中查看。 + +### Auto 一定比手动选择更好吗？ + +不一定。Auto 适合没有明确模型偏好、希望减少选择成本的场景。如果你已经知道要使用哪个模型，直接指定会更可控。 + +### 使用 Auto 会更贵吗？ + +Auto 本身不额外收费，但不同模型的单价和 Token 消耗可能不同。请在 **用量信息（Usage）** 页面查看每次请求的实际费用。 diff --git a/i18n/zh-Hans/docusaurus-plugin-content-docs/current/llmservice/pricing-and-usage.md b/i18n/zh-Hans/docusaurus-plugin-content-docs/current/llmservice/pricing-and-usage.md index bd85993..e0fbbea 100644 --- a/i18n/zh-Hans/docusaurus-plugin-content-docs/current/llmservice/pricing-and-usage.md +++ b/i18n/zh-Hans/docusaurus-plugin-content-docs/current/llmservice/pricing-and-usage.md @@ -62,6 +62,10 @@ 文档价格为 B.AI 平台模型标准参考价，仅供基础计费说明使用。B.AI 可能会通过充值赠送及账户权益等方式，为用户提供更低的实际使用成本。具体价格、赠送积分及账户权益请以平台页面展示及最终账单为准。 ::: +:::info Auto 模式计费说明 +Auto 不是单独计费的模型。每次请求将按照实际执行的模型、Token 用量和工具使用情况结算；最终费用请以对话用量详情或 **用量信息（Usage）** 页面记录为准。详见 [Auto 模式](./auto-mode.md)。 +::: + ### 缓存定价说明 - **缓存写入（Cache Write）：** 当 token 首次写入 Prompt Cache 时产生的消耗。大多数提供方不会额外加价，通常与标准输入价格一致。Claude 模型和 GPT-5.6 模型的缓存写入会在对应输入价格基础上增加 25%；GPT-5.6 长上下文缓存写入价格按对应长上下文输入档位计算。 diff --git a/i18n/zh-Hans/docusaurus-plugin-content-docs/current/llmservice/quick-start.md b/i18n/zh-Hans/docusaurus-plugin-content-docs/current/llmservice/quick-start.md index 8658aee..47612d7 100644 --- a/i18n/zh-Hans/docusaurus-plugin-content-docs/current/llmservice/quick-start.md +++ b/i18n/zh-Hans/docusaurus-plugin-content-docs/current/llmservice/quick-start.md @@ -48,7 +48,7 @@ LLM Service 采用去中心化登录方式，支持通过 Web3 钱包进行授 -- **选择模型：** 点击当前模型名称，从列表中选择需要使用的模型（如 GPT、Claude、Gemini 等） +- **选择模型：** 点击当前模型名称，从列表中选择需要使用的模型（如 GPT、Claude、Gemini 等）。如果暂时不确定该选择哪一款模型，可选择 **Auto**，由 B.AI 为每次请求匹配合适的可用模型。详见 [Auto 模式](./auto-mode.md)。 - **发送消息：** 在输入框输入内容，点击发送按钮或按 Enter @@ -79,4 +79,3 @@ LLM Service 采用积分（Credits）计费，通过链上支付获取。 完成以上步骤后，你即可使用多种主流 AI 模型。 - diff --git a/package.json b/package.json index 433201a..1d95f26 100644 --- a/package.json +++ b/package.json @@ -1,6 +1,6 @@ { "name": "@x402-tron/docs", - "version": "1.3.11", + "version": "1.3.12", "description": "x402-tron documentation", "license": "MIT", "resolutions": { diff --git a/sidebars.js b/sidebars.js index 9499b74..3440ad8 100644 --- a/sidebars.js +++ b/sidebars.js @@ -138,6 +138,7 @@ const sidebars = { items: [ { type: 'doc', id: 'llmservice/introduction', label: 'Introduction' }, { type: 'doc', id: 'llmservice/quick-start', label: 'Quick Start' }, + { type: 'doc', id: 'llmservice/auto-mode', label: 'Auto Mode' }, { type: 'doc', id: 'llmservice/pricing-and-usage', label: 'Pricing and Usage' }, { type: 'category',