WORLDGATE API · OPENAI-COMPATIBLE GATEWAY

WorldGate API for cheap GPT & Claude.
One key. Lower cost.

WorldGate API provides transparent per-token pricing for GPT and Claude models, with prepaid usage, streaming, and transparent token-level billing.

Get an API Key Compare OpenRouter

Built on infrastructure developers already use

Cloudflare
Supabase
Vercel
GitHub

MODEL DIRECTORY / LIVE

One key.
Every model you need.

14

Currently listed GPT and Claude models routed through one OpenAI-compatible endpoint.

Explore every model
01
G
GPT 5.5GPT family / Flagship reasoning
OpenAI-compatible$0.2600 / 1M inputPUBLISHED RATE
02
C
Claude Sonnet 5Claude family / Coding & automation
OpenAI-compatible$0.1040 / 1M inputPUBLISHED RATE
03
C
Claude Opus 5Claude family / Premium reasoning
OpenAI-compatible$0.2600 / 1M inputPUBLISHED RATE
04
C
Claude Opus 4.8Claude family / Deep reasoning
OpenAI-compatible$0.2600 / 1M inputPUBLISHED RATE
04
G
GPT 5.4 MiniGPT family / Economical throughput
OpenAI-compatible$0.0390 / 1M inputPUBLISHED RATE

WHY WORLDGATE EXISTS

Great models should not be locked behind a big budget.

Developers everywhere have ambitious ideas, but access to top AI models can still be limited by regional availability, high minimums, and unpredictable API bills.

WorldGate combines demand across our network and routes every request through the most efficient available provider. That lets us offer the models you already know at a much lower price, through one simple API.

Our principle Less spent on access. More spent on building.
Many providers

We continuously compare capacity, region, and price.

One WorldGate API

A single endpoint handles routing and model access.

Lower model cost

Current WorldGate rates and the referenced provider prices are shown together for a clear comparison.

ONE-LINE MIGRATION

Keep your SDK.
Change one URL.

WorldGate speaks the API format your app already understands. Point your existing client to our gateway and start routing across models immediately.

01 OpenAI-compatible 02 Streaming supported 03 Automatic failover
quickstart.jsJAVASCRIPT / SDK
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.WORLDGATE_API_KEY,
  baseURL: "https://api.worldgateapi.com/v1",
});

const response = await client.chat.completions.create({
  model: "gpt-5.5",
  messages: [{ role: "user", content: "Ship it." }],
});
200 OK342msRouted via Singapore

MODEL PRICING

Built for
lower inference cost.

Prices are published per 1M input, output, and cache-read tokens, with no hidden request fees.

Claude Models

USD per 1M tokens · US
ModelInputOutputCache RCache WOfficial I/ODiscount
Claude Opus 5$0.2600$1.56$0.0260$0.3250——
Claude Fable 5$3.25$9.10$0.3250$4.06——
Claude Opus 4.8$0.2600$1.56$0.0260$0.3250——
Claude Opus 4.7$0.2600$1.56$0.0260$0.3250——
Claude Opus 4.6$0.2600$1.56$0.0260$0.3250——
Claude Sonnet 5$0.1040$0.5200$0.0104$0.1300——
Claude Sonnet 4.6$0.1560$0.7800$0.0156$0.1950——
Claude Haiku 4.5$0.0520$0.2600$0.0052$0.0650——
Claude usage is billed per 1M tokens at the rates shown above.Start with Claude

GPT Models

USD per 1M tokens · US
ModelInputOutputCache RCache WOfficial I/ODiscount
GPT 5.6 Sol$0.2600$1.56$0.0260$0.3250——
GPT 5.5$0.2600$1.56$0.0260$0.3250——
GPT 5.6 Terra$0.1300$0.7800$0.0130$0.1625——
GPT 5.4$0.1300$0.7800$0.0130$0.1625——
GPT 5.6 Luna$0.0520$0.3120$0.0052$0.0650——
GPT 5.3 Codex Spark$0.0910$0.7280$0.0091$0.1138——
GPT 5.4 Mini$0.0390$0.2340$0.0039$0.0488——
GPT usage is billed per 1M tokens at the rates shown above.Start with GPT

Kimi Models

USD per 1M tokens · US
ModelInputOutputCache RCache WOfficial I/ODiscount
Kimi K3$2.21$11.05$0.2210$2.76——
Kimi K2.7 Code$0.7800$3.25$0.0780$0.9750——
Kimi K2.6$0.7800$3.25$0.0780$0.9750——
Kimi K2.5$0.5200$2.60$0.0520$0.6500——
Kimi usage is billed per 1M tokens at the rates shown above.Start with Kimi

GLM Models

USD per 1M tokens · US
ModelInputOutputCache RCache WOfficial I/ODiscount
GLM 5.2$0.9100$3.25$0.0910$1.14——
GLM 5.1$0.7800$2.60$0.0780$0.9750——
GLM usage is billed per 1M tokens at the rates shown above.Start with GLM

DeepSeek Models

USD per 1M tokens · US
ModelInputOutputCache RCache WOfficial I/ODiscount
DeepSeek V4 Pro$0.3510$0.7800$0.0351$0.4387——
DeepSeek V4 Flash$0.1040$0.2600$0.0104$0.1300——
DeepSeek usage is billed per 1M tokens at the rates shown above.Start with DeepSeek

Qwen Models

USD per 1M tokens · US
ModelInputOutputCache RCache WOfficial I/ODiscount
Qwen 3.7 Max$1.56$3.51$0.1300$1.63——
Qwen 3.7 Plus$0.2210$1.04$0.0221$0.2762——
Qwen usage is billed per 1M tokens at the rates shown above.Start with Qwen

MiniMax Models

USD per 1M tokens · US
ModelInputOutputCache RCache WOfficial I/ODiscount
MiniMax M2.7$0.2210$1.04$0.0221$0.2762——
MiniMax M3$0.2210$1.04$0.0221$0.2762——
MiniMax usage is billed per 1M tokens at the rates shown above.Start with MiniMax

CLAUDE API GUIDES

Find the right Claude API price for your workload.

Compare Claude API pricing, cheap Opus 5 access, and Claude Code setup before you add production traffic.

Claude API pricingPer-token rates and prepaid billing OpenAI-compatible APIMigration and endpoint guide Claude Code API setupInstall, configure, and troubleshoot Claude API PythonCompatible SDK tutorial Claude API Node.jsJavaScript and TypeScript LangChain integrationCustom endpoint and spend controls AI API cost calculatorEstimate input and output spend Cheap Claude Opus 5$0.2600 input / $1.56 output guide Cheap Claude API keyOne key for Opus, Sonnet, Haiku Cheap Claude subscriptionsPrepaid API alternative Claude API modelsChoose by quality and cost OpenRouter alternativeFocused GPT and Claude gateway

NETWORK INTELLIGENCE

The fastest path
to every model.

WorldGate automatically routes requests based on latency, provider health, pricing, and availability.

View Live Network
User
WG
AI
Claude Opus 5 Healthy
29ms
G
GPT 5.6 Sol Healthy
23ms
C
Claude Sonnet 5 Healthy
24ms
G
GPT 5.4 Healthy
31ms

BUILT IN THE REAL WORLD

From first prototype
to millions of tokens.

One infrastructure layer for teams that need to experiment quickly, control spend, and scale without rewriting their stack.

01
SOLO DEVELOPERS

Launch the idea before the credits run out.

Try listed models with a small balance, compare output quality, and monitor usage for every API key.

No monthly subscription
02
AI PRODUCTS

Route every task to the model that fits.

Use economical models for routine work and stronger reasoning models only when the request actually needs them.

14 listed models
03
AGENCIES & TEAMS

See where model spend goes.

Separate keys by client or environment and monitor token-level usage with a request ledger.

Per-key usage

COMMON QUESTIONS

Before you send your first request.

Clear answers about pricing, reliability, compatibility, and how WorldGate works.

Ask us directly
How can WorldGate offer models for less?

We aggregate usage, negotiate capacity, and route requests across multiple providers. The savings from that infrastructure are reflected directly in your usage price.

Which GPT and Claude models are currently listed?

The current catalogue includes six GPT-family models and eight Claude-family models. Supported identifiers and current token prices are published in the model directory.

What happens if a provider goes down?

Health checks detect provider issues and automatically route eligible requests to another available endpoint, reducing interruptions without requiring changes to your integration.

Can I use my existing OpenAI SDK?

Yes. Our API follows the familiar OpenAI-compatible format, so most projects only need a new base URL and WorldGate API key.

Is there a subscription or minimum spend?

No. You can start with prepaid usage and scale at your own pace. There is no monthly commitment for the standard plan.