大模型 API：AI 工具与模型排行榜

AI/ML API (AIMLAPI) is a prepaid API service, contracting through Boiler Labs FZ-LLC in the UAE and also listing AIMLAPI OÜ in Estonia, that sells text, image, video and audio models from many makers through one key. Its current prices for OpenAI, Anthropic and Google models sit about 30% above the makers' list prices.

最适合：

Developers who want one prepaid key for text, image, video and audio models; Claude Code, Codex and OpenCode users who want to switch providers without editing config files

最大短板：

Prices above the makers' list prices.

APIMODELS 是一家按量付费的 AI API 中转站，用一个 Key 提供约 150 个模型，价格自定、多数低于厂商官方价，省下的利润投在备用渠道上：近 30 天去掉内容审核拒绝后，请求成功率 98.62%。APIMODELS 是我们自己的产品，和 GenGrades 出自同一团队。

最适合：

想用一次 $10 充值、一张账单调用多种模型的个人开发者和小团队; 上游出问题一小时也不能让生成队列停下来的产品

最大短板：

首页宣传的折扣比它自己的对比表大。

Chutes is a decentralized inference network on Bittensor subnet 64, run by Chutes Global Corp, where anonymous GPU operators serve open-weight models such as DeepSeek, Kimi and GLM through an OpenAI-compatible API. It ended its free tier on March 15, 2026, runs all 14 public models in trusted execution environments, and relies on TAO token subsidies.

最适合：

Roleplay users on Janitor AI or SillyTavern who want cheap open models; Hobby coders running OpenCode, Aider or Cline on open models

最大短板：

No free tier, and prices have risen.

Cloudflare AI Gateway is Cloudflare's free proxy for AI requests, adding logging, caching, rate limits and fallbacks in front of 24 providers. Since August 2025 its Unified Billing lets you buy Cloudflare credits and call OpenAI, Anthropic, Google, xAI and Groq models without your own keys, at list prices plus a 5% fee on every credit purchase.

最适合：

Teams already building on Cloudflare Workers or Workers AI; Developers who want logs, caching and spend limits in front of their own OpenAI or Anthropic keys

最大短板：

A 5% fee on every credit purchase.

CometAPI is a prepaid API service that sells text, image, video and audio models through one key at, by its own rule, 80% of the makers' list prices. Its terms and privacy pages name no company and use Hong Kong law, and it says it reaches the low prices through bulk credit buying and routing.

最适合：

Price-driven developers who want Claude, GPT and Gemini below list price; Hobby projects mixing chat models with Midjourney, Sora 2 or Kling

最大短板：

No named company.

Eden AI is a Lyon-based API gateway, run by Eden AI SAS, that puts about 500 language models and 250 'expert' models for OCR, speech, translation and image work behind one key and one prepaid balance. It passes provider prices through and, since January 2026, adds a 5.5% platform fee on self-serve credits.

最适合：

EU companies that need a DPA and EU-only processing for AI calls; Teams that need OCR, document parsing, speech or translation alongside LLMs

最大短板：

A 5.5% fee on top of list prices.

Inworld Router is Inworld AI's model-routing API, which sends OpenAI- and Anthropic-style requests to third-party models at the providers' list prices with no markup and can route by user metadata. On October 14, 2026 Inworld folded it into Inworld Inference, which also serves open models on Inworld's own hardware; inworld.ai/router now redirects there.

最适合：

Consumer apps that want to test models per user segment, language or plan; Voice and character apps already using Inworld's text-to-speech or speech-to-text

最大短板：

The product was renamed and refocused within days.

Kilo Gateway is the OpenAI-compatible model API behind the Kilo Code agent, run by Kilo Code, Inc., which Anaconda acquired in July 2026. It bills tokens at the upstream rate plus a 5% fee on credit purchases, and most of its 400 models match OpenRouter's IDs and prices.

最适合：

Kilo Code users in VS Code, JetBrains, the CLI or Kilo's cloud agents who want one balance; Developers with Z.ai, Kimi Code or OpenCode Go plans who want them next to pay-as-you-go models

最大短板：

It is largely a layer on OpenRouter.

NanoGPT is a pay-per-prompt AI service run by NanoGPT LLC, a U.S. company, with a web chat and an OpenAI- and Anthropic-compatible API for about 670 models at the makers' list prices. You can start with 0.10 USD in crypto and no account, or pay 12 USD a month for a plan built around open-weight models.

最适合：

Privacy-minded individuals who want pay-per-prompt access without a subscription or an account; SillyTavern, JanitorAI and other roleplay users on a budget

最大短板：

Tool calling for agents.

Novita AI is an inference host for open-weight models such as DeepSeek, GLM, Kimi and Qwen, with image, video, GPU and sandbox services; it does not resell GPT, Claude or Gemini. Its per-token prices match or undercut the makers' own, its terms promise zero data retention by default, and it does not name the company behind it.

最适合：

Developers who want new DeepSeek, GLM, Kimi and Qwen releases without hosting them; Teams that need a zero-retention host for open-weight models

最大短板：

No closed models.

OpenCode Zen, called OpenCode Console in the docs since October 2026, is the pay-as-you-go model gateway from Anomaly Innovations, the team behind the OpenCode coding agent. It sells about 80 coding models near the makers' list prices plus a 4.6% + 0.31 USD card fee; a Console breach that month may have exposed some users' saved provider keys.

最适合：

OpenCode users who want one prepaid balance for closed and open coding models; Claude Code, Codex, Pi or Hermes users who want a second source for the same models

最大短板：

A Console breach in October 2026.

OpenRouter is a New York company's API relay that sells access to more than 680 models from over 100 hosting providers through one key, at each provider's list price plus a 5.5% fee on credit purchases. Stripe agreed to buy it on August 19, 2026; no closing had been announced by October 21, 2026.

最适合：

Developers who want to try or switch between many models from one key and one bill; Teams running open-weight models such as DeepSeek, Kimi or GLM who want to choose among competing hosts

最大短板：

The same model can behave differently depending on the host.

Poe API is Quora's developer API for the bots in its Poe chat app, opened on July 31, 2025, with OpenAI-compatible endpoints and an Anthropic-compatible one for official Claude bots. Calls spend the same points as a Poe subscription, so there is no separate API bill; current Claude bots cost about 1% more than Anthropic's list price.

最适合：

Poe subscribers who want to spend their points in Claude Code, Cline, Roo Code or Cursor; SillyTavern and other chat front-end users who already pay for Poe

最大短板：

Points are hard to predict, and they are the main complaint.

Requesty is a London-based LLM gateway, run by Requesty Ltd and seed-funded by 20VC in 2025, that sells OpenAI- and Anthropic-compatible access to more than 600 models at list price plus a 5% top-up fee. It stands out for routing policies, spend controls and an EU endpoint in Frankfurt.

最适合：

European teams that want their gateway to process requests inside the EU; Engineering teams running Claude Code, Cline, Roo Code or Codex who want spend caps per developer

最大短板：

Prompts and outputs are logged by default.

TypeSafe AI 是旧金山的一家 AI 实验室，首个模型 Jev 不生成文字，而是返回带概率的类型化决策，供软件直接使用。Jev 于 2026 年 9 月 15 日开放早期访问，同时宣布 4000 万美元种子轮；彭博社 10 月 9 日报道，它又以 75 亿美元估值融资约 8.7 亿美元。

最适合：

要在软件里对大量文本做分类、打分或分流的开发者; 用 LLM 做筛选或内容审核、单次调用成本不断累积的团队

最大短板：

它不是文本模型。

Vercel AI Gateway is Vercel's hosted API for about 480 text, image, video and speech models behind one key, billed from prepaid credits at the providers' list prices with no markup, including when you bring your own keys. It became generally available on August 21, 2025, and gives each Vercel team 5 USD of free credit a month.

最适合：

Developers already deploying on Vercel or building with the AI SDK; Teams running coding agents such as Claude Code, Codex or Cursor who want one bill and spend tracking

最大短板：

Billing bugs have blocked paying accounts.

ZenMux is an LLM gateway, operated under its terms by NexaMind Singapore Pte. Ltd., that sells Claude, GPT, Gemini and many Chinese models through OpenAI, Anthropic and Vertex-style endpoints at the makers' list token prices. It adds automatic credits for bad or slow answers, which it calls insurance, and 20 to 200 USD monthly coding plans.

最适合：

Developers using Claude Code, Codex, Cursor or Gemini CLI who want one key for Western and Chinese models; Teams that want provider pinning and fallbacks across Anthropic, Vertex AI and Bedrock

最大短板：

Claude access has broken at the upstream.