Opus 5, Sonnet 5 and GPT-5.6 through one key. Pay per token as you go, no subscription. Works with Cursor, Claude Code, SDKs and agents.
Pay only for what you use: per 1M tokens, per model
No subscriptions. No rate-limit games. Just tokens in, tokens out.
No packages or prepaid plans. You pay per 1M tokens for each model you use.
Works with Anthropic and OpenAI-compatible SDKs, editors, agents, and automation.
Join our Telegram channel and get a dollar on your balance. Enough to check latency and compatibility before you pay anything.
Requests pass through our API supplier. Processing locations and data limitations are published in the legal section. Do not send personal or confidential data before reviewing those terms.
Explain the architecture of this project.
Start with the entry points and follow the data flow.
Read the app routes and API client
The client sends requests to the API using your key.
Keep the key in an environment variable, outside the repository.
Choose a model and check its available token rates.
Everything you need to know before getting started.
Top up from $1 and connect the API. Pay only for what you actually use.
GET API ACCESSCreate an account in a minute: username and password, or Google, GitHub and Telegram.
Add balance via card, SBP, crypto, or Platega.
Generate an API key from your dashboard.
Set the base URL in your tool and start building.
Use named keys to account for team and project usage. Track spend per key against one shared account balance and budget; separate keys do not create independent quotas.
Name keys by team, project, or integration and use per-key statistics for usage accounting.
The account budget covers shared usage. It does not provide a separate monthly cap for each team or key.
Choose Premium or Cheap, allow Premium fallback for Cheap if needed, set expiry, and track usage per key.
Pay-as-you-go per model. Real per-token rates and compatibility across your stack.
Each card shows which models are available, what they are good at and how much they cost.
Direct access to the models, every feature, very cheap caching. Usually the cheapest for IDE work.
Claude Opus 5.5 is a text model by Anthropic. Context: 1,000,000 tokens. Supports reasoning. Accepts images.
Claude Opus 5 is a text model by Anthropic. Context: 1,000,000 tokens. Supports reasoning. Accepts images.
Claude Fable 5.1 is a text model by Anthropic. Context: 1,000,000 tokens. Supports reasoning. Accepts images.
Claude Fable 5 is a text model by Anthropic. Context: 1,000,000 tokens. Supports reasoning. Accepts images.
Claude Sonnet 5 is a text model by Anthropic. Context: 1,000,000 tokens. Supports reasoning. Accepts images.
Claude Opus 4.8 is a text model by Anthropic. Context: 1,000,000 tokens. Supports reasoning. Accepts images.
Claude Opus 4.7 is a text model by Anthropic. Context: 1,000,000 tokens. Supports reasoning. Accepts images.
Claude Opus 4.6 is a text model by Anthropic. Context: 1,000,000 tokens. Supports reasoning. Accepts images.
Claude Sonnet 4.6 is a text model by Anthropic. Context: 1,000,000 tokens. Supports reasoning. Accepts images.
GPT-6 Astra is a text model by OpenAI. Context: 1,000,000 tokens. Supports reasoning. Accepts images.
GPT-6.1 Sol is a text model by OpenAI. Context: 1,000,000 tokens. Supports reasoning. Accepts images.
GPT-6 Sol is a text model by OpenAI. Context: 1,000,000 tokens. Supports reasoning. Accepts images.
GPT-6 Luna is a text model by OpenAI. Context: 1,000,000 tokens. Supports reasoning.
GPT-5.6 Sol is a text model by OpenAI. Context: 1,000,000 tokens. Supports reasoning. Accepts images.
GPT-5.6 Terra is a text model by OpenAI. Context: 1,000,000 tokens. Supports reasoning. Accepts images.
GPT-5.6 Luna is a text model by OpenAI. Context: 1,000,000 tokens. Supports reasoning.
Gemini 3.8 Flash is a text model by Google. Context: 1,000,000 tokens. Supports reasoning.
Gemini 3.7 Flash is a text model by Google. Context: 1,000,000 tokens. Supports reasoning.
Gemini 3.1 Pro is a text model by Google. Context: 1,000,000 tokens. Supports reasoning. Accepts images.
Qwen 3.8 Max is a text model by Alibaba. Context: 1,000,000 tokens. Supports reasoning. Accepts images.
Qwen 3.8 Flash is a text model by Alibaba. Context: 1,000,000 tokens. Supports reasoning.
DeepSeek V4.1 Flash is a text model by DeepSeek. Context: 1,000,000 tokens. Supports reasoning.
DeepSeek V4 Pro is a text model by DeepSeek. Context: 1,000,000 tokens. Supports reasoning. Accepts images.
DeepSeek V4 Flash is a text model by DeepSeek. Context: 1,000,000 tokens. Supports reasoning.
GLM 5.3 is a text model by Zhipu. Context: 1,000,000 tokens. Supports reasoning.
GLM 5.3 Flash is a text model by Zhipu. Context: 1,000,000 tokens. Supports reasoning.
GLM 5.2 is a text model by Zhipu. Context: 1,000,000 tokens. Supports reasoning. Accepts images.
Grok 4.7 is a text model by xAI. Context: 256,000 tokens. Accepts images.
Grok 4.6 is a text model by xAI. Context: 256,000 tokens. Accepts images.
Grok 4.5 is a text model by xAI. Context: 256,000 tokens. Accepts images.
GPT Image 2.5 is an image generation model by OpenAI.
GPT Image 2 is an image generation model by OpenAI.
Nano Banana 2 is an image generation model by Google.
Nano Banana Pro is an image generation model by Google.
Nano Banana is an image generation model by Google.
Veo 3.1 is a video generation model by Google.
Wan 3.0 is a video generation model by Alibaba.
Hailuo 02 is a video generation model by MiniMax.
Grok Imagine is a video generation model by xAI.
Seedance 2.5 is a video generation model by ByteDance.
Seedance 2.0 is a video generation model by ByteDance.
Seedance 1 is a video generation model by ByteDance.
Kling 3.0 is a video generation model by Kuaishou.
Kling 2.6 is a video generation model by Kuaishou.
PixVerse V6 is a video generation model by PixVerse.
Runway Gen-4.5 is a video generation model by Runway.
Runway Gen-4 is a video generation model by Runway.
Estimate your spend before you top up
200K in, 30K out, on Claude Fable 5
Presets are typical volumes, not a promise. Put your own numbers in; rates are pulled live from our pricing.
OpenAI & Anthropic compatible
