KIE$0.03per image
−85.8%vs Official
Official$0.211per image
透明定价，为规模化而生
以更低成本接入领先 AI 模型。一个透明钱包，无隐藏费用，失败生成不收费。
平均节省
最高 88%
失败任务
$0
隐藏费用
无
Model
Category
KIE — vs — Official
GPT Image 2 1K
Image
KIE$0.03per image
−85.8%vs
Official$0.211per image
GPT Image 2 2K
Image
KIE$0.05per image
−88.3%vs
Official$0.428per image
GPT Image 2 4K
Image
KIE$0.08per image
−80.0%vs
Official$0.40per image
GPT Image 2.5 1K
Image
KIE$0.03per image
−43.1%vs
Official$0.053per image
GPT Image 2.5 2K
Image
KIE$0.05per image
−53.3%vs
Official$0.107per image
Google Nano Banana 2 1K
Image
KIE$0.04per image
−40.3%vs
Official$0.067per image
Google Nano Banana 2 2K
Image
KIE$0.06per image
−40.6%vs
Official$0.101per image
Google Nano Banana 2 4K
Image
KIE$0.09per image
−40.4%vs
Official$0.151per image
Google Nano Banana Pro 1K/2K
Image
KIE$0.09per image
−32.8%vs
Official$0.134per image
Google Nano Banana Pro 4K
Image
KIE$0.12per image
−50.0%vs
Official$0.24per image
Grok Imagine Video 1.5 Image-to-Video 720p
Video
KIE$0.0225per sec
−71.9%vs
Official$0.08per sec
Grok Imagine Video 1.5 Image-to-Video 480p
Video
KIE$0.012per sec
−85.0%vs
Official$0.08per sec
Grok Imagine Video 720p
Video
KIE$0.0225per sec
−55.0%vs
Official$0.05per sec
Grok Imagine Video 480p
Video
KIE$0.012per sec
−76.0%vs
Official$0.05per sec
Google Veo 3.1 Lite 1080p
Video
KIE$0.175per video
−72.7%vs
Official$0.64per video
Google Veo 3.1 Fast 1080p
Video
KIE$0.325per video
−66.1%vs
Official$0.96per video
Google Veo 3.1 Quality 1080p
Video
KIE$1.275per video
−60.2%vs
Official$3.20per video
Google Veo 3.1 Quality Image-to-Video 4K
Video
KIE$1.85per video
−61.5%vs
Official$4.80per video
GPT-6 Astra Input
LLM
KIE$2.80per 1M tokens
−72.0%vs
Official$10.00per 1M tokens
GPT-6 Astra Output
LLM
KIE$14.00per 1M tokens
−72.0%vs
Official$50.00per 1M tokens
GPT-6 Sol Input
LLM
KIE$0.60per 1M tokens
−70.0%vs
Official$2.00per 1M tokens
GPT-6 Sol Output
LLM
KIE$3.00per 1M tokens
−70.0%vs
Official$10.00per 1M tokens
Claude Opus 5.5 Input
LLM
KIE$1.60per 1M tokens
−60.0%vs
Official$4.00per 1M tokens
Claude Opus 5.5 Output
LLM
KIE$8.00per 1M tokens
−60.0%vs
Official$20.00per 1M tokens
Gemini 3.8 Flash Input
LLM
KIE$0.225per 1M tokens
−70.0%vs
Official$0.75per 1M tokens
Gemini 3.8 Flash Output
LLM
KIE$1.125per 1M tokens
−70.0%vs
Official$3.75per 1M tokens
KIE$0.05per image
−88.3%vs Official
Official$0.428per image
KIE$0.08per image
−80.0%vs Official
Official$0.40per image
KIE$0.03per image
−43.1%vs Official
Official$0.053per image
KIE$0.05per image
−53.3%vs Official
Official$0.107per image
KIE$0.04per image
−40.3%vs Official
Official$0.067per image
KIE$0.06per image
−40.6%vs Official
Official$0.101per image
KIE$0.09per image
−40.4%vs Official
Official$0.151per image
KIE$0.09per image
−32.8%vs Official
Official$0.134per image
KIE$0.12per image
−50.0%vs Official
Official$0.24per image
KIE$0.0225per sec
−71.9%vs Official
Official$0.08per sec
KIE$0.012per sec
−85.0%vs Official
Official$0.08per sec
KIE$0.0225per sec
−55.0%vs Official
Official$0.05per sec
KIE$0.012per sec
−76.0%vs Official
Official$0.05per sec
KIE$0.175per video
−72.7%vs Official
Official$0.64per video
KIE$0.325per video
−66.1%vs Official
Official$0.96per video
KIE$1.275per video
−60.2%vs Official
Official$3.20per video
KIE$1.85per video
−61.5%vs Official
Official$4.80per video
KIE$2.80per 1M tokens
−72.0%vs Official
Official$10.00per 1M tokens
KIE$14.00per 1M tokens
−72.0%vs Official
Official$50.00per 1M tokens
KIE$0.60per 1M tokens
−70.0%vs Official
Official$2.00per 1M tokens
KIE$3.00per 1M tokens
−70.0%vs Official
Official$10.00per 1M tokens
KIE$1.60per 1M tokens
−60.0%vs Official
Official$4.00per 1M tokens
KIE$8.00per 1M tokens
−60.0%vs Official
Official$20.00per 1M tokens
KIE$0.225per 1M tokens
−70.0%vs Official
Official$0.75per 1M tokens
KIE$1.125per 1M tokens
−70.0%vs Official
Official$3.75per 1M tokens
KIE 精选模型
直接通过我们的 API 使用最受欢迎、表现力最强的一批模型。
VIDSeedance 2.5
$0.085 / sSeedance 2.5 is ByteDance's next-gen video model, generating up to 30-second 4K clips in a single take with up to 50 multimodal references and precise region-level editing.
Online试用模型
IMGGPT Image 2.5
$0.03 / imgGPT Image 2.5 is OpenAI's latest image model, delivering sharper detail, more natural lighting, and precise multi-turn editing that keeps people and products consistent — about 50% faster than GPT Image 2.
Online试用模型
IMGGPT Image 2
$0.03 / imgGPT Image 2 is OpenAI’s next-gen image model built for stronger photorealism, cleaner image editing, sharper text rendering, and more polished product photography.
Online试用模型
IMGNano Banana 2
$0.04 / imgMeet Nano Banana 2, Google’s Gemini 3.1 Flash Image model, now available via Kie AI API. Built for developers, it combines lightning-fast speed with Pro-level quality, accurate text rendering, strong character consistency, and scalable image generation and editing workflows.
Online试用模型
VIDVeo 3.1
$0.15 / videoGoogle DeepMind’s upgraded AI video model for realistic motion generation, extended clip duration, multi-image reference control, and synchronized audio output in native 1080p.
Online试用模型
VID
Grok Imagine Video 1.5
$0.012 / sGrok Imagine Video 1.5 is xAI's latest image-to-video model, turning a single image into smooth, expressive motion at up to 720p — built for fast, high-volume creative workflows.
Online试用模型
VID
Kling 3.0
$0.07 / sKling 3.0 is Kling AI’s video generation model that creates videos from text and images, supports multi-shot storytelling, and produces native audio with cinematic control up to 15 seconds.
Online试用模型
VIDSeedance 2.5
$0.085 / sSeedance 2.5 is ByteDance's next-gen video model, generating up to 30-second 4K clips in a single take with up to 50 multimodal references and precise region-level editing.
Online试用模型
IMGGPT Image 2.5
$0.03 / imgGPT Image 2.5 is OpenAI's latest image model, delivering sharper detail, more natural lighting, and precise multi-turn editing that keeps people and products consistent — about 50% faster than GPT Image 2.
Online试用模型
IMGGPT Image 2
$0.03 / imgGPT Image 2 is OpenAI’s next-gen image model built for stronger photorealism, cleaner image editing, sharper text rendering, and more polished product photography.
Online试用模型
IMGNano Banana 2
$0.04 / imgMeet Nano Banana 2, Google’s Gemini 3.1 Flash Image model, now available via Kie AI API. Built for developers, it combines lightning-fast speed with Pro-level quality, accurate text rendering, strong character consistency, and scalable image generation and editing workflows.
Online试用模型
VIDVeo 3.1
$0.15 / videoGoogle DeepMind’s upgraded AI video model for realistic motion generation, extended clip duration, multi-image reference control, and synchronized audio output in native 1080p.
Online试用模型
VID
Grok Imagine Video 1.5
$0.012 / sGrok Imagine Video 1.5 is xAI's latest image-to-video model, turning a single image into smooth, expressive motion at up to 720p — built for fast, high-volume creative workflows.
Online试用模型
VID
Kling 3.0
$0.07 / sKling 3.0 is Kling AI’s video generation model that creates videos from text and images, supports multi-shot storytelling, and produces native audio with cinematic control up to 15 seconds.
Online试用模型
VIDSeedance 2.5
$0.085 / sSeedance 2.5 is ByteDance's next-gen video model, generating up to 30-second 4K clips in a single take with up to 50 multimodal references and precise region-level editing.
Online试用模型
IMGGPT Image 2.5
$0.03 / imgGPT Image 2.5 is OpenAI's latest image model, delivering sharper detail, more natural lighting, and precise multi-turn editing that keeps people and products consistent — about 50% faster than GPT Image 2.
Online试用模型
IMGGPT Image 2
$0.03 / imgGPT Image 2 is OpenAI’s next-gen image model built for stronger photorealism, cleaner image editing, sharper text rendering, and more polished product photography.
Online试用模型
IMGNano Banana 2
$0.04 / imgMeet Nano Banana 2, Google’s Gemini 3.1 Flash Image model, now available via Kie AI API. Built for developers, it combines lightning-fast speed with Pro-level quality, accurate text rendering, strong character consistency, and scalable image generation and editing workflows.
Online试用模型
VIDVeo 3.1
$0.15 / videoGoogle DeepMind’s upgraded AI video model for realistic motion generation, extended clip duration, multi-image reference control, and synchronized audio output in native 1080p.
Online试用模型
VID
Grok Imagine Video 1.5
$0.012 / sGrok Imagine Video 1.5 is xAI's latest image-to-video model, turning a single image into smooth, expressive motion at up to 720p — built for fast, high-volume creative workflows.
Online试用模型
VID
Kling 3.0
$0.07 / sKling 3.0 is Kling AI’s video generation model that creates videos from text and images, supports multi-shot storytelling, and produces native audio with cinematic control up to 15 seconds.
Online试用模型
VIDWan 3.0
$0.04 / sWan 3.0 is Alibaba's omni-reference video model for longer storytelling — up to 30-second videos from text, images, video or even documents, with precise editing and immersive audio.
Online试用模型
IMGSeedream 5.0 Pro
$0.035 / imgSeedream 5.0 Pro is ByteDance's design-aware image model — turning dense text and data into professional layouts, with pixel-level interactive editing and more realistic light, material and skin.
Online试用模型
AUDSuno V6
$0.06 / reqSuno V6 is Suno's newest music model — full vocal and instrumental tracks with richer realism, natural-language song editing, mashups and sampling, plus V6 Mini for speed and V6 Wild for bolder results.
Online试用模型
AUDGemini 3.8 Flash TTS
$0.35 / MGemini 3.8 Flash TTS is Google's latest expressive speech model — 100+ languages, 2,000+ ready-made voices, voice cloning from a 30-second sample, and native two-speaker dialogue.
Online试用模型
LLMGPT-6 Astra
$2.80 / MGPT-6 Astra is OpenAI's flagship model — its smartest and best-aligned yet, built for advanced reasoning, coding, research, computer use and long-horizon agentic work.
Online试用模型
LLMClaude Opus 5.5
$1.60 / MClaude Opus 5.5 is Anthropic's newest frontier model — Fable-class intelligence at lower cost and 30%+ faster output than Opus 5, excelling at agentic coding, computer use and knowledge work.
Online试用模型
LLMGemini 3.8 Flash
$0.225 / MGemini 3.8 Flash is Google's latest Flash model — fast, low-cost multimodal intelligence for high-volume chat, coding and agentic workloads.
Online试用模型
VIDWan 3.0
$0.04 / sWan 3.0 is Alibaba's omni-reference video model for longer storytelling — up to 30-second videos from text, images, video or even documents, with precise editing and immersive audio.
Online试用模型
IMGSeedream 5.0 Pro
$0.035 / imgSeedream 5.0 Pro is ByteDance's design-aware image model — turning dense text and data into professional layouts, with pixel-level interactive editing and more realistic light, material and skin.
Online试用模型
AUDSuno V6
$0.06 / reqSuno V6 is Suno's newest music model — full vocal and instrumental tracks with richer realism, natural-language song editing, mashups and sampling, plus V6 Mini for speed and V6 Wild for bolder results.
Online试用模型
AUDGemini 3.8 Flash TTS
$0.35 / MGemini 3.8 Flash TTS is Google's latest expressive speech model — 100+ languages, 2,000+ ready-made voices, voice cloning from a 30-second sample, and native two-speaker dialogue.
Online试用模型
LLMGPT-6 Astra
$2.80 / MGPT-6 Astra is OpenAI's flagship model — its smartest and best-aligned yet, built for advanced reasoning, coding, research, computer use and long-horizon agentic work.
Online试用模型
LLMClaude Opus 5.5
$1.60 / MClaude Opus 5.5 is Anthropic's newest frontier model — Fable-class intelligence at lower cost and 30%+ faster output than Opus 5, excelling at agentic coding, computer use and knowledge work.
Online试用模型
LLMGemini 3.8 Flash
$0.225 / MGemini 3.8 Flash is Google's latest Flash model — fast, low-cost multimodal intelligence for high-volume chat, coding and agentic workloads.
Online试用模型
VIDWan 3.0
$0.04 / sWan 3.0 is Alibaba's omni-reference video model for longer storytelling — up to 30-second videos from text, images, video or even documents, with precise editing and immersive audio.
Online试用模型
IMGSeedream 5.0 Pro
$0.035 / imgSeedream 5.0 Pro is ByteDance's design-aware image model — turning dense text and data into professional layouts, with pixel-level interactive editing and more realistic light, material and skin.
Online试用模型
AUDSuno V6
$0.06 / reqSuno V6 is Suno's newest music model — full vocal and instrumental tracks with richer realism, natural-language song editing, mashups and sampling, plus V6 Mini for speed and V6 Wild for bolder results.
Online试用模型
AUDGemini 3.8 Flash TTS
$0.35 / MGemini 3.8 Flash TTS is Google's latest expressive speech model — 100+ languages, 2,000+ ready-made voices, voice cloning from a 30-second sample, and native two-speaker dialogue.
Online试用模型
LLMGPT-6 Astra
$2.80 / MGPT-6 Astra is OpenAI's flagship model — its smartest and best-aligned yet, built for advanced reasoning, coding, research, computer use and long-horizon agentic work.
Online试用模型
LLMClaude Opus 5.5
$1.60 / MClaude Opus 5.5 is Anthropic's newest frontier model — Fable-class intelligence at lower cost and 30%+ faster output than Opus 5, excelling at agentic coding, computer use and knowledge work.
Online试用模型
LLMGemini 3.8 Flash
$0.225 / MGemini 3.8 Flash is Google's latest Flash model — fast, low-cost multimodal intelligence for high-volume chat, coding and agentic workloads.
Online试用模型
面向开发者的集成体验
通过统一、直观的 API，几分钟内即可完成接入。
STEP 01
选择模型
使用统一参数结构，从顶级媒体模型中选择，无需学习不同平台的 API。
VIDIMGAUDLLMVIDLLMIMGAUDIMGVIDLLMAUD
STEP 02
发送异步请求
提交生成任务。所有模型共用稳定可预测的 API 网关，切换模型非常轻松。
API
STEP 03
Webhook 与轮询
提供 webhook URL，在媒体生成完成时即时收到通知，或按架构需求使用标准状态轮询。
WEBHOOK
ReadyPOLL
For coding agents
让你的 coding agent 直接调用 KIE
安装一次 KIE skills。Claude Code、Codex、Cursor 等 agent 就能自己找模型、读参数、发任务并取回结果，不用再把文档贴进对话。
npx skills add https://kie.ai适用于Claude CodeCodexCursorGrok Build
› Make a 1:1 product photo of a white ceramic mug with GPT Image 2 and save it to ./public
0sSearching KIE models · 213 available
1sReading parameters · gpt-image-2-text-to-image
2sTask submitted · c7f3c3ad…
65sGenerated · 3 credits
66sSaved ./public/mug.png
kie-models
200+ 模型用图像、视频、音乐和语音模型构建应用。你的 agent 会选择模型、核对价格、上传输入文件，并轮询直到结果就绪。
kie-chat-agents
配置指南用你的 KIE credits 在 KIE 聊天模型上运行 Codex、Claude Code 或 Grok Build。覆盖每个 agent 的 base URL、密钥和模型名。
为什么开发者选择 KIE
我们抽象掉多平台生成的复杂性，让你专注于构建自己的应用。
01
Lower-Cost Access
更低成本接入领先模型
KIE 通过灵活的积分系统提供更低成本的领先 AI 模型访问。相比官方 API 和其他市场，许多模型价格约低 30%，部分热门模型可节省 60–70%。
02
Unified API
一个 API 接入 100+ AI 模型
通过一个统一 API 访问视频、图像、音频和 LLM 模型。无需重建后端，即可在 Veo、Kling、Seedance、Runway、Claude、GPT、Gemini、Nano Banana、Suno 等模型间切换。
03
Reliability
99% 可用性，24/7 监控，稳定可靠
KIE 通过 API 可用性监控、全天候任务监控、异步任务追踪、webhook 回调和智能备用路由，保障高并发 AI 生成流程稳定运行。
04
Playground
Playground 中免费试用 AI API
在编写生产代码前，可直接在 Playground 测试 KIE API。尝试不同提示词、比较模型输出、调整参数，并在真实 API 环境中评估视频、图像、音频和 LLM 模型。
05
Drop-in
简单的 AI API 集成
一次集成即可访问视频、图像、音频和 LLM 领域的领先模型。清晰文档、统一任务流程和一致 API 模式，让团队只需修改 model_id 即可切换模型。
06
Security
可靠的数据安全
安全的 API 认证、加密请求和受控任务处理，帮助保护你的提示词、素材和生成结果。
07
Support
全年私有支持渠道
KIE 全年在线支持生产级 AI 工作流。API 用户可通过 Telegram 和 Discord 的专属私有渠道获得一对一支持，处理 API 错误、任务日志、webhook 回调、账单和集成问题。
技术团队看重 KIE 的什么
受到下一代 AI 原生应用构建者的信赖。
testimonials / 01
Production我们已经基于 KIE API 部署了基础设施。对 Veo 和 Kling 生成使用 webhook 回调，而不是长轮询，为我们节省了大量服务器成本。
常见问题
关于 KIE 集成与计费，你需要了解的一切。
KIE.ai 是一个统一的 AI API 平台，让开发者可以通过一套一致的 API 接入领先的视频、图像、音频和 LLM 模型。
你不需要逐一集成不同供应商，而是可以使用 KIE 以更简单的流程、更清晰的文档和更具竞争力的价格连接多个顶级 AI 模型。在许多情况下，KIE 的模型接入成本低于官方定价，通常低约 30%，部分模型甚至更具成本优势。
KIE 面向需要在生产环境中稳定、可扩展且高性价比接入 AI 模型的团队构建。
KIE 支持多个类别的领先 AI 模型，包括视频、图像、音频和 LLM。
视频生成方面，KIE 支持 Seedance 2.0、Veo、Kling、Grok Imagine 等模型。
图像生成方面，KIE 支持 GPT Image 2、Nano Banana 以及其他高级图像模型。
音频生成方面，KIE 支持 Suno、ElevenLabs 以及其他音频相关 API。
LLM 方面，KIE 支持 Claude、GPT、Gemini 等领先语言模型。
你可以在模型市场查看完整模型列表：https://kie.ai/market
你可以在 KIE 定价页面查看当前模型价格：https://kie.ai/pricing
不同模型可能使用不同计费单位，例如按图片、按秒、按视频或按百万 tokens 计费。定价页面会展示每个模型适用的计费方式。
KIE 的价格公开透明。你可以在这里查看计费和充值详情：https://kie.ai/billing
对于较大金额充值，KIE 可能提供赠送 credits。例如，充值 $1,250 会额外获得 10% 赠送 credits，使实际使用成本约比标价低 10%。
每个模型页面都包含 playground，你可以在集成 API 前测试模型。
新用户还会获得 80 个免费 credits 用于测试，因此你可以在正式开发前尝试不同模型和工作流。
你可以在这里查看 API 文档：https://docs.kie.ai/
文档包含 API endpoints、请求示例、参数、callback 用法、响应格式以及特定模型的集成细节。
可以。安装 KIE skills 后，你的 coding agent 就能自己找模型、读参数、发任务并下载结果。你也可以用 KIE credits 在 KIE 聊天模型上运行 Claude Code、Codex 或 Grok Build。
npx skills add https://kie.ai不会。失败任务不收费。
如果生成任务失败，KIE 不会扣除该失败任务的 credits。这能帮助开发者更放心地测试和运行生产工作负载，无需为未成功的生成付费。
KIE 为用户提供专门的技术支持渠道。
你可以通过这个页面加入一对一支持群：https://kie.ai/vip-support
在那里你可以选择通过 Discord 或 Telegram 加入。我们的支持团队可以协助 API 集成、任务错误、callback 问题、计费问题和生产使用问题。
是的。KIE 设计用于支持生产级用例和商业工作流。
我们 24/7 监控模型可用性和任务性能，并持续优化路由和后端资源以提升稳定性。当出现问题时，团队会尽快定位并解决。
不会。KIE credits 不会过期。
你的 credits 会保留在账户中，可随时使用。
可以。KIE 支持为企业用户开具发票。
完成充值后，你可以在计费页面手动生成发票：https://kie.ai/billing
公司信息可以在开票流程中添加。
你可以在 Logs 中查看已提交任务和生成结果：
https://kie.ai/logs
Logs 页面可以查看任务记录、状态、参数、结果以及相关任务详情。
登录后，你可以在 Updates 页面查看最近的模型发布、价格更新和功能变化：
https://kie.ai/api-updates
KIE 不对运行中的任务设置严格并发限制。
对于新请求提交，默认限制是每 10 秒 20 个新请求，也就是每分钟最多 120 个新请求。这对大多数用户和生产用例已经足够。
如果你收到 429 rate limit 错误，可以联系 KIE 支持申请更高的速率限制。我们会根据你的使用量和业务需求进行评估后调整。
媒体文件，包括图片、视频和音频，会保留 14 天。
日志数据会保留 2 个月。
超过这些保留周期后，数据将被永久删除。请提前下载并备份重要文件或任务日志。
立即使用全球顶级 AI 模型开始构建
在一个平台中访问视频、图像、音乐和聊天 API，更快、更实惠，也更适合开发者。选择 Veo 3.1 API、Runway Aleph API、Suno API 等模型，均具备稳定性能和简单集成体验。
获取免费 API Key