系统文档

请求示例

curl https://your-domain.example/v1/models \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY"

响应示例

{
  "object": "list",
  "data": [
    {
      "id": "gpt-image-2",
      "object": "model",
      "created": 0,
      "owned_by": "gpt2image"
    }
  ]
}

请求字段

返回与流式

实现说明

• 本站当前只实现模型列表，不实现 /v1/models/{model} 详情。
• 返回模型会按套餐过滤；Ultra 用户可见更多 Responses 模型。

请求示例

curl https://your-domain.example/v1/credits \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY"

响应示例

{
  "object": "credit_balance",
  "account": {
    "balance": 15702.45,
    "total_earned": 20000,
    "total_spent": 4297.55,
    "status": "active"
  },
  "api_key": {
    "credit_limit": 1000,
    "credits_used": 12.7,
    "credits_remaining": 987.3,
    "unlimited": false
  }
}

请求字段

返回与流式

实现说明

• API Key 限额只限制该 Key 自身；走本站平台计费路径时仍必须有足够账户积分。
• 命中用户自接 API 时不扣本站账户积分，也不增加 Key 已用额度。
• 生成失败退款、审核拦截结算和实际尺寸后修正会同步修正 Key 已用额度。

请求示例

# 1. 普通对话生图；默认返回 URL，content 中会追加 Markdown 图片链接
curl https://your-domain.example/v1/chat/completions \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.4",
    "image_model": "gpt-image-2",
    "messages": [
      { "role": "system", "content": "你是专业视觉海报设计师。" },
      { "role": "user", "content": "生成一张科技企业宣传海报，16:9，蓝白配色" }
    ],
    "size": "1536x864",
    "quality": "high",
    "response_format": "url"
  }'

# 2. 多模态输入，image_url 会作为本轮真实参考图输入
curl https://your-domain.example/v1/chat/completions \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.4-mini",
    "image_model": "gpt-image-2",
    "messages": [
      {
        "role": "user",
        "content": [
          { "type": "text", "text": "参考这张产品图，生成一张电商主图" },
          { "type": "image_url", "image_url": { "url": "https://example.com/product.png" } }
        ]
      }
    ],
    "size": "1024x1024",
    "response_format": "url"
  }'

# 3. 流式返回；文本走 chat.completion.chunk，自定义 partial_image 事件返回流式预览
curl -N https://your-domain.example/v1/chat/completions \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -H "Accept: text/event-stream" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.4",
    "messages": [
      { "role": "user", "content": "生成一张未来城市概念图" }
    ],
    "size": "1024x1024",
    "stream": true
  }'

响应示例

{
  "id": "chatcmpl_...",
  "object": "chat.completion",
  "created": 1713833628,
  "model": "gpt-5.4",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "已生成图片。\n\n![generated image 1](https://your-domain.example/api/storage/generations/...)",
        "images": [
          {
            "url": "https://your-domain.example/api/storage/generations/...",
            "revised_prompt": "...",
            "generation_id": "gen_..."
          }
        ]
      },
      "finish_reason": "stop"
    }
  ],
  "images": [
    {
      "url": "https://your-domain.example/api/storage/generations/...",
      "generation_id": "gen_..."
    }
  ],
  "generation_id": "gen_...",
  "generationId": "gen_...",
  "credits_consumed": 2.31,
  "usage": null
}

# stream=true 时的 SSE 片段
data: {"id":"chatcmpl_...","object":"chat.completion.chunk","choices":[{"index":0,"delta":{"content":"正在生成..."},"finish_reason":null}]}

event: chat.completion.partial_image
data: {"type":"chat.completion.partial_image","index":0,"partial_image_index":0,"url":"https://your-domain.example/api/storage/generations/..."}

data: {"id":"chatcmpl_...","object":"chat.completion.chunk","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"generation_id":"gen_...","credits_consumed":2.31}

请求字段

返回与流式

实现说明

• 上游 API 配置可选择 Chat Completions 上游模式：默认 responses 会请求上游 /responses + image_generation tool，保留图片生成、图片引用和 Codex/Responses 能力。
• 选择 chat_completions 后，本站 /v1/chat/completions 会请求命中上游的 /chat/completions；这更适合纯聊天兼容，但是否能返回图片取决于上游实现。Agent 和 /v1/responses 不受该配置影响。
• OpenAI 官方 Chat Completions 并不定义“生成图片”的标准返回字段；本站为了兼容对话生图，在 Chat Completions 外形上扩展 choices[].message.images、顶层 images，并在 content 中追加 Markdown 图片链接。严格按官方生图协议接入时，建议使用 /v1/images/generations、/v1/images/edits 或 /v1/responses。
• 该接口走页面 Chat 的非 Agent 模式，不会注入 web_search、continue_generation，也不会展示 Agent 多轮任务卡。
• 调度类型是 chat，可命中 Web 账号、Codex/Responses 账号或支持 /responses 的外接 API 后端；用户自接 API 可用时仍保持最高优先级。
• 计费等同页面 Chat：先收 Chat 每轮基础积分，再按最终图片实际尺寸和数量追加图片积分、审核积分和分组倍率。

请求示例

# 1. 官方 Images 风格，默认返回 b64_json
curl https://your-domain.example/v1/images/generations \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2",
    "prompt": "A cute baby sea otter",
    "n": 1,
    "size": "1024x1024",
    "quality": "medium",
    "moderation": "auto",
    "background": "auto"
  }'

# 2. 返回 URL，并关闭本站提示词优化
curl https://your-domain.example/v1/images/generations \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-1.5",
    "prompt": "一张赛博朋克城市夜景，雨后霓虹反光",
    "n": 2,
    "size": "1024x1024",
    "quality": "high",
    "moderation": "low",
    "response_format": "url",
    "output_format": "webp",
    "output_compression": 85,
    "background": "transparent",
    "prompt_optimization": false
  }'

# 3. Codex/Responses 后端专用参数；普通 Images API 后端可能忽略
curl https://your-domain.example/v1/images/generations \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2",
    "prompt": "生成一张 16:9 产品海报",
    "size": "1536x864",
    "response_format": "url",
    "output_format": "jpeg",
    "output_compression": 90,
    "gptModel": "gpt-5.4",
    "thinking": "high",
    "promptOptimization": false
  }'

# 4. mixed 分组按可配置像素区间强制调度 Web 账号；非 mixed 分组会忽略 force_web
curl https://your-domain.example/v1/images/generations \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2",
    "prompt": "一张 1:1 头像海报",
    "size": "1024x1024",
    "response_format": "url",
    "force_web": true
  }'

# 5. 流式返回；也可用 Accept: text/event-stream 触发
curl -N https://your-domain.example/v1/images/generations \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -H "Accept: text/event-stream" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2",
    "prompt": "一张透明玻璃材质的未来感咖啡杯",
    "size": "1024x1024",
    "response_format": "url",
    "stream": true
  }'

# 6. 异步模式；也可在 URL 后追加 ?async=true。callback_url 可选
curl https://your-domain.example/v1/images/generations \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-1.5",
    "prompt": "一张透明背景的产品图标",
    "size": "1024x1024",
    "response_format": "url",
    "output_format": "png",
    "background": "transparent",
    "async": true,
    "callback_url": "https://your-server.example/callback"
  }'

响应示例

{
  "created": 1713833628,
  "data": [
    {
      "url": "https://your-domain.example/api/storage/generations/...",
      "revised_prompt": "..."
    }
  ],
  "generation_id": "gen_...",
  "generationId": "gen_...",
  "credits_consumed": 1.31,
  "usage": null
}

# stream=true 时的 SSE 片段
event: image_generation.partial_image
data: {"type":"image_generation.partial_image","index":0,"partial_image_index":0,"url":"https://your-domain.example/api/storage/generations/..."}

event: image_generation.completed
data: {"type":"image_generation.completed","index":0,"generation_id":"...","generationId":"...","model":"gpt-image-2","size":"1024x1024","credits_consumed":1.31,"url":"https://your-domain.example/api/storage/generations/...","data":[{"url":"https://your-domain.example/api/storage/generations/...","revised_prompt":"..."}]}

# async=true 的立即响应
{
  "id": "task_...",
  "object": "image.generation",
  "model": "gpt-image-2",
  "status": "processing",
  "created": 1713833628,
  "created_at": "2026-05-28T00:00:00.000Z",
  "generation_id": "gen_..."
}

# 查询任务
curl https://your-domain.example/v1/images/task_... \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY"

# 完成后的任务响应或回调 payload
{
  "id": "task_...",
  "object": "image",
  "model": "gpt-image-2",
  "status": "completed",
  "created": 1713833628,
  "created_at": "2026-05-28T00:00:00.000Z",
  "completed": 1713833700,
  "completed_at": "2026-05-28T00:01:12.000Z",
  "data": [{"url": "https://your-domain.example/api/storage/generations/..."}],
  "generation_id": "gen_...",
  "generationId": "gen_...",
  "credits_consumed": 1.31,
  "usage": null
}

请求字段

返回与流式

实现说明

• 该接口不会调用页面 /api/images/generate，而是直接进入共享 service 层。
• 如果命中 Responses 账号池，内部会把图片请求转换成 Responses image_generation tool 请求。
• n/count 批量张数属于一次 HTTP 请求；一次 10 张会创建 10 条生成记录并按 10 张计费。运行时按套餐的生图并发受限并行，超过并发上限的图片会在本批次内排队等待。
• 并发与排队：底层只有一条进程内生图队列，任务按套餐队列优先级排序，同优先级先进先出；队列同时受全局并发和单用户生图并发限制。全局并发可在后台「系统设置 > 模型 > 全局生图并发」配置，环境变量 IMAGE_GENERATION_GLOBAL_CONCURRENCY 只作为兜底默认值。批量请求额外有请求内 runner，只启动套餐允许的并发数，剩余图片留在本批次内等待，不会一次性塞满底层队列。
• 排队等待阶段不会创建 generation，也不会扣图像生成积分；底层队列排队超过 IMAGE_GENERATION_QUEUE_TIMEOUT_MS 会返回 429 类错误。单张任务开始执行后才进入 20 分钟运行超时，运行超时按失败结算规则处理积分。
• Web 后端无法严格控制输出尺寸和输出格式；本站保存时会按实际图片头识别扩展名和 MIME。
• background=transparent 并非所有模型都支持；OpenAI 官方文档当前列出 gpt-image-1.5、gpt-image-1、gpt-image-1-mini 支持透明背景，且通常还要求 png 或 webp 输出。不支持的上游可能直接返回 HTTP 400，而不是自动降级。
• async 任务当前为进程内状态，30 分钟后过期；服务重启或多实例切换会导致未完成任务无法继续查询，callback 已发送的结果不受影响。
• 如果实际生成尺寸与请求尺寸不一致，本站会按检测到的实际尺寸修正记录和计费。
• 官方 Images API 可能返回 usage；本站当前 usage 通常为 null，但会通过顶层 credits_consumed、错误对象或流式完成事件返回本站结算积分。命中用户自接 API 时不扣本站积分。

请求示例

# 1. multipart 上传参考图
curl https://your-domain.example/v1/images/edits \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -F model="gpt-image-2" \
  -F prompt="把参考图改成电影海报风格" \
  -F n="1" \
  -F size="1024x1024" \
  -F quality="high" \
  -F moderation="auto" \
  -F response_format="url" \
  -F output_format="jpeg" \
  -F output_compression="90" \
  -F background="opaque" \
  -F 'image[]=@/path/to/reference.png'

# 2. multipart 多参考图 + mask + Codex/Responses 参数
curl https://your-domain.example/v1/images/edits \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -F model="gpt-image-2" \
  -F prompt="只重绘 mask 区域，保持人物脸部不变" \
  -F size="1536x1024" \
  -F quality="medium" \
  -F response_format="b64_json" \
  -F promptOptimization="false" \
  -F gpt_model="gpt-5.4" \
  -F thinking="medium" \
  -F 'image[]=@/path/to/person.png' \
  -F 'image_2=@/path/to/style.png' \
  -F mask="@/path/to/mask.png"

# 3. JSON 图片 URL；推荐 images，image_url/image_urls 只是兼容快捷字段
curl https://your-domain.example/v1/images/edits \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2",
    "prompt": "把参考图改成干净的电商主图",
    "images": [
      "https://example.com/reference.png",
      { "image_url": "https://example.com/detail.webp" }
    ],
    "image_url": "https://example.com/single-reference.png",
    "image_urls": ["https://example.com/extra.jpg"],
    "mask_url": "https://example.com/mask.png",
    "mask_image_url": "https://example.com/mask-alt.png",
    "n": 1,
    "size": "1024x1024",
    "quality": "auto",
    "moderation": "low",
    "response_format": "url",
    "output_format": "webp",
    "output_compression": 80,
    "background": "transparent",
    "prompt_optimization": false,
    "gptModel": "gpt-5.4-mini",
    "thinking": "low"
  }'

# 4. mixed 分组按可配置像素区间强制调度 Web 账号；非 mixed 分组会忽略 force_web
curl https://your-domain.example/v1/images/edits \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2",
    "prompt": "保留人物，改成电影剧照质感",
    "images": ["https://example.com/reference.png"],
    "size": "1024x1024",
    "response_format": "url",
    "force_web": true
  }'

# 5. 流式图生图
curl -N https://your-domain.example/v1/images/edits \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -H "Accept: text/event-stream" \
  -F model="gpt-image-2" \
  -F prompt="保留构图，改成水彩插画风格" \
  -F size="1024x1024" \
  -F response_format="url" \
  -F stream="true" \
  -F 'image=@/path/to/reference.png'

# 6. 异步图生图；也可在 URL 后追加 ?async=true。callback_url 可选
curl https://your-domain.example/v1/images/edits \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -F model="gpt-image-1.5" \
  -F prompt="去除背景，输出透明 PNG" \
  -F size="1024x1024" \
  -F response_format="url" \
  -F output_format="png" \
  -F background="transparent" \
  -F async="true" \
  -F callback_url="https://your-server.example/callback" \
  -F 'image=@/path/to/reference.png'

响应示例

{
  "created": 1713833628,
  "data": [
    {
      "url": "https://your-domain.example/api/storage/generations/...",
      "revised_prompt": "..."
    }
  ],
  "generation_id": "gen_...",
  "generationId": "gen_...",
  "credits_consumed": 1.31,
  "usage": null
}

# stream=true 时的 SSE 片段
event: image_edit.partial_image
data: {"type":"image_edit.partial_image","index":0,"partial_image_index":0,"url":"https://your-domain.example/api/storage/generations/..."}

event: image_edit.completed
data: {"type":"image_edit.completed","index":0,"generation_id":"...","generationId":"...","model":"gpt-image-2","size":"1024x1024","credits_consumed":1.31,"url":"https://your-domain.example/api/storage/generations/...","data":[{"url":"https://your-domain.example/api/storage/generations/...","revised_prompt":"..."}]}

# async=true 的任务查询和回调响应格式同 /v1/images/generations

请求字段

返回与流式

实现说明

• URL 图片会先由本站服务端下载并校验公网可访问性、类型和大小。
• 不支持私网、localhost、metadata/internal 域名或带用户名密码的 URL。
• 官方 JSON file_id 图片引用当前未实现，请使用公网 image_url 或 multipart 上传。
• background=transparent 并非所有模型都支持；OpenAI 官方文档当前列出 gpt-image-1.5、gpt-image-1、gpt-image-1-mini 支持透明背景，且通常还要求 png 或 webp 输出。不支持的上游可能直接返回 HTTP 400，而不是自动降级。
• async 任务当前为进程内状态，30 分钟后过期；服务重启或多实例切换会导致未完成任务无法继续查询，callback 已发送的结果不受影响。

请求示例

# 1. JSON Agent 生图；默认返回 URL。默认需要 Ultra，可在能力矩阵 externalApi.agent 调整。
curl https://your-domain.example/v1/agents/images \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.4",
    "image_model": "gpt-image-2",
    "prompt": "联网查询浙江双元科技公开资料，迭代生成一张企业宣传海报",
    "size": "1536x1024",
    "quality": "high",
    "thinking": "medium",
    "agent_max_rounds": 3,
    "agent_force_max_rounds": false,
    "response_format": "url"
  }'

# 2. 带参考图 URL。images / image_url / image_urls 会合并去重。
curl https://your-domain.example/v1/agents/images \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.4-mini",
    "image_model": "gpt-image-2",
    "prompt": "参考这张产品图，先分析卖点，再生成一张电商海报",
    "images": ["https://example.com/product.png"],
    "size": "1024x1024",
    "agent_max_rounds": 2
  }'

# 3. multipart 上传参考图和 PDF/文本附件。
curl https://your-domain.example/v1/agents/images \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -F model="gpt-5.4" \
  -F image_model="gpt-image-2" \
  -F prompt="阅读附件资料并生成一张展会宣传海报" \
  -F size="1536x1024" \
  -F response_format="url" \
  -F agent_max_rounds="3" \
  -F 'image[]=@/path/to/reference.png' \
  -F 'file=@/path/to/company-profile.pdf'

# 4. 流式 Agent。会持续返回 agent.event / agent.partial_image / agent.completed。
curl -N https://your-domain.example/v1/agents/images \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -H "Accept: text/event-stream" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.4",
    "image_model": "gpt-image-2",
    "prompt": "先搜索资料，再迭代生成一张科技蓝企业海报",
    "size": "1536x1024",
    "stream": true,
    "agent_max_rounds": 2,
    "agent_force_max_rounds": true
  }'

响应示例

{
  "object": "agent.image_run",
  "created": 1713833628,
  "generation_id": "gen_...",
  "generationId": "gen_...",
  "model": "gpt-5.4",
  "size": "1536x1024",
  "response_text": "已完成资料检索并生成海报。",
  "agent_round_count": 2,
  "credits_consumed": 8.42,
  "data": [
    {
      "url": "https://your-domain.example/api/storage/generations/...",
      "revised_prompt": "...",
      "output_role": "agent_draft"
    },
    {
      "url": "https://your-domain.example/api/storage/generations/...",
      "revised_prompt": "...",
      "output_role": "final"
    }
  ],
  "agent_events": [],
  "usage": null
}

# stream=true 时的 SSE 片段
event: agent.event
data: {"type":"agent.event","event":{"kind":"web_search","status":"completed","title":"联网搜索完成","detail":"浙江双元科技 官网"}}

event: agent.partial_image
data: {"type":"agent.partial_image","partial_image_index":0,"url":"https://your-domain.example/api/storage/generations/..."}

event: agent.completed
data: {"type":"agent.completed","generation_id":"...","generationId":"...","agent_round_count":2,"credits_consumed":8.42,"data":[{"url":"https://your-domain.example/api/storage/generations/...","output_role":"final"}]}

请求字段

返回与流式

实现说明

• 该接口是本站扩展，不是 OpenAI 官方接口；/api/v1/agents/images 是同一 handler 的别名。
• 默认要求 Ultra 套餐；管理员可在套餐能力矩阵中调整 externalApi.agent。
• 该接口强制 requiresResponsesBackend，不会命中 Web 账号；支持 Codex/Responses 账号或支持 /responses 的外接 API 后端。
• 不会调用页面 /api/images/chat；它和页面 Agent 共享 runImageGenerationForUser service 层。
• generate_image_batch 并发工具暂未开放，避免破坏线性迭代和 Responses 粘性会话。

请求示例

# 1. 最小 Responses 生图请求；需要 Pro 套餐
curl https://your-domain.example/v1/responses \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.4",
    "input": "生成一张 1:1 的未来感产品渲染图",
    "size": "1024x1024",
    "quality": "high",
    "moderation": "auto"
  }'

# 2. 显式 image_generation tool，并指定图片模型
curl https://your-domain.example/v1/responses \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.4",
    "input": "生成一张横版科技产品 KV",
    "tools": [{ "type": "image_generation", "model": "gpt-image-2" }],
    "size": "1536x864",
    "quality": "medium",
    "reasoning": { "effort": "low" },
    "store": true
  }'

# 3. 带参考图的 Responses 输入
curl https://your-domain.example/v1/responses \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.4-mini",
    "input": [
      {
        "role": "user",
        "content": [
          { "type": "input_text", "text": "参考这张图，换成冬季海报风格" },
          { "type": "input_image", "image_url": "https://example.com/reference.png" }
        ]
      }
    ],
    "tools": [{ "type": "image_generation", "model": "gpt-image-2" }],
    "size": "1024x1024",
    "output_format": "webp",
    "output_compression": 85,
    "moderation": "low"
  }'

# 4. 续接上一轮，并使用流式返回
curl -N https://your-domain.example/v1/responses \
  -H "Authorization: Bearer $GPT2IMAGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.4",
    "previous_response_id": "resp_previous_id",
    "input": "在上一张图基础上加一个月亮",
    "tools": [{ "type": "image_generation", "model": "gpt-image-2" }],
    "size": "1024x1024",
    "reasoning": { "effort": "minimal" },
    "stream": true
  }'

响应示例

{
  "id": "resp_...",
  "object": "response",
  "created_at": 1713833628,
  "status": "completed",
  "model": "gpt-5.4",
  "output": [
    {
      "id": "ig_...",
      "type": "image_generation_call",
      "status": "completed",
      "result": "..."
    }
  ],
  "usage": null,
  "metadata": {
    "generation_id": "...",
    "credits_consumed": 1.31,
    "size": "1024x1024"
  }
}

# stream=true 时的 SSE 片段
event: response.output_item.done
data: {"type":"response.output_item.done","item":{"id":"ig_...","type":"image_generation_call","status":"completed","result":"..."}}

event: response.completed
data: {"type":"response.completed","response":{"id":"resp_...","object":"response","created_at":1713833628,"status":"completed","model":"gpt-5.4","output":[{"id":"ig_...","type":"image_generation_call","status":"completed","result":"..."}],"usage":null,"metadata":{"generation_id":"...","credits_consumed":1.31,"size":"1024x1024"}}}

请求字段

返回与流式

实现说明

• 该接口需要专业版或更高套餐。
• 该接口不是 Chat Completions；普通对话生图请使用 /v1/chat/completions，Responses 工具语义请使用本接口。
• input_image 只支持 image_url/data URL；file_id/file 输入当前不会作为参考图使用。
• 显式传 tools 但不包含 image_generation 会返回错误，避免模型只产出文本而不生图。
• 页面 Chat 模式只提供普通多模态对话/生图语义；Agent 模式默认提供 image_generation、web_search 和线性续跑工具 continue_generation，不强制 tool_choice，模型按任务自行选择工具。
• 页面 Chat/Agent 支持上传文本/代码类本地文件作为上下文读取；不会读取用户在提示词中写入的服务器本地路径。
• 页面 Chat/Agent 的每轮基础积分由后台「套餐能力矩阵」按套餐配置，默认 Chat 每轮 1 积分、Agent 每轮 3 积分；生成图片时再按实际尺寸和输出数量追加图片积分。
• Agent 会把上一轮文字、工具结果和已生成图片喂回下一轮，让模型自行判断是否继续改版；最大轮数由系统设置 IMAGE_AGENT_MAX_ROUNDS 控制，默认 3。当前没有接入 generate_image_batch 这类并发批量工具，以免打散 Responses 粘性会话。
• Agent 多轮产生的 image_generation_call 会作为自动迭代版本展示，最后一张作为默认选中版本。