一個驗證，對應所有 provider
一組 RunAPI key 即可解鎖整個 catalog。無需分開帳戶，亦無需為每個整合輪換密鑰。
curl -X POST https://runapi.ai/api/v1/gpt_image_2/text_to_image \
-H "Authorization: Bearer $RUNAPI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "gpt-image-2",
"prompt": "Generate a greeting card with '新年快乐 Happy New Year' in decorative calligraphy, gold and red theme."
}'import { GptImage2Client } from "@runapi.ai/gpt-image-2";
const client = new GptImage2Client();
const result = await client.textToImage.run({
model: "gpt-image-2",
prompt: "Generate a greeting card with '新年快乐 Happy New Year' in decorative calligraphy, gold and red theme.",
});<?php
require __DIR__ . "/vendor/autoload.php";
use RunApi\GptImage2\GptImage2Client;
$client = new GptImage2Client();
$result = $client->textToImage->run([
'model' => 'gpt-image-2',
'prompt' => 'Generate a greeting card with \'新年快乐 Happy New Year\' in decorative calligraphy, gold and red theme.',
]);require "runapi/gpt_image_2"
client = RunApi::GptImage2::Client.new
result = client.text_to_image.run(
model: "gpt-image-2",
prompt: "Generate a greeting card with '新年快乐 Happy New Year' in decorative calligraphy, gold and red theme."
)npx skills add runapi-ai/gpt-image-2 -g# Claude Code
claude mcp add runapi -s user -- npx -y @runapi.ai/mcp
# Codex
codex plugin install runapi-mcp@agents
# Cursor / Windsurf / VS Code
npx @runapi.ai/mcp init cursorOVERVIEW
關於 GPT Image 2
GPT Image 2 is OpenAI's latest image generation model featuring near-perfect multilingual text rendering within generated images. It improves on instruction following and fine-grained visual detail over its predecessor.
供應商
OpenAI
模式
Image
PRICING
收費
失敗的生成不會收費
Text to image
$0.06-$0.16
/ image
Output resolution: 1k
$0.06
Output resolution: 2k
$0.10
Output resolution: 4k
$0.16
Edit image
$0.16
/ image
規格說明
技術詳情
| 模型 ID | gpt-image-2 |
| 供應商 | OpenAI |
| 模式 | image |
| 任務類型 | asynchronous |
| 計費單位 | call |
| API endpoint | /api/v1/gpt_image_2/text_to_image |
| 商業授權 | 是 — 已透過 API 包含 |
| 目錄狀態 | 可正常運作 |
Agent 整合
透過 Agent 執行 GPT Image 2
運作方式
開始使用 GPT Image 2
-
選擇模型
挑選符合輸出類型、品質門檻與延遲目標的模型與變體。
-
設定
設定 RunAPI key，並在 coding workspace 安裝 model skill。
-
開發
使用 skill 指引，在你的 app 內加入模型功能。
-
接收
透過 task ID 查詢、在支援時串流，或處理 webhook callback。
為何透過 RunAPI 使用 GPT Image 2 API
統一定價與計費
按次以 USD 收費，每月結算。失敗的生成不會收費。
內含 schema 的 skill
型別化 schema 與 setup 備註打包在 model skill 內，讓實作從正確契約開始。
GPT Image 2 常見問題
How accurate is the multilingual text rendering?
GPT Image 2 可以渲染多種語言嘅文字——英文同拉丁字母準確率大約 99%，中文、日文、韓文、阿拉伯文等非拉丁字母約 95% 以上。
What is the maximum output resolution?
Up to 4K resolution (in beta); standard output is 2K. Both text-to-image and image editing endpoints support high-resolution output.
What is Thinking mode for image generation?
Thinking mode lets the model reason about composition and layout before generating, improving accuracy for complex multi-element scenes and precise text placement.
How does GPT Image 2 handle text in non-Latin scripts?
It renders CJK characters, Arabic, Cyrillic, and other scripts with near-native accuracy — a significant improvement over prior image generation models.
Can I generate and edit in the same workflow?
Yes — generate from text, then refine with image editing. Both endpoints share the same API shape.
我應該先用邊個版本？
先選最平而且符合你質素要求的版本。大多數團隊會先用快速版本，之後再升級到 pro 用於正式生產。
相似模型