Standard Compute gives you access to multiple model tiers through a single API. Use simple aliases to pick the right balance of speed and capability for each workflow, and let our routing handle the rest.
Standard Compute organizes its available models into three tiers: Base, Mid, and Top. Each tier represents a different class of model capability and is mapped to your plan so you always get the best performance your subscription allows.
When you send a request you can specify a tier alias ("fast", "standard", or "premium") in the model field, or you can use a specific model name. The API will route your request to the best available model within the tier your plan supports.
The Base tier includes GPT-3.5-class models optimized for speed and efficiency. These models handle straightforward tasks like text classification, simple summarization, data extraction, and template-based generation with very low latency.
Use the alias "fast" in the model field to target this tier. Base tier models are available on all plans, including the Starter plan. They are ideal for high-volume automation workflows where speed matters more than reasoning depth.
Typical response times for Base tier models are 200 to 500 milliseconds for short completions, making them well-suited for real-time automation chains in n8n, Make, and Zapier.
The Mid tier includes GPT-4o-mini-class models that strike a balance between capability and speed. These models excel at multi-step reasoning, nuanced text generation, code completion, and structured data transformation.
Use the alias "standard" in the model field to target this tier. Mid tier models are available on the SMB plan and above. They are the recommended default for most automation workflows that require reliable, high-quality output.
Mid tier models typically respond in 500 milliseconds to 2 seconds depending on the length of the completion. They support longer context windows than Base tier models, allowing you to pass in more context from previous workflow steps.
The Top tier includes the most capable models available, comparable to GPT-4o and Claude Sonnet. These models are best for complex reasoning, detailed content generation, code analysis, strategic planning, and any task where output quality is the top priority.
Use the alias "premium" in the model field to target this tier. Top tier models are available on the Agency plan. They provide the highest accuracy and the most coherent long-form output.
Response times for Top tier models range from 1 to 5 seconds depending on the prompt length and max_tokens setting. Because these models require more compute, they have a slightly lower per-minute rate limit, but usage remains unlimited within your plan.
Standard Compute uses intelligent model routing to maximize uptime and minimize latency. When you send a request with a tier alias, the system selects the best available model within that tier based on current load, provider availability, and your geographic region.
This means you do not need to worry about individual provider outages. If one upstream provider experiences issues, the router automatically fails over to an alternative model within the same quality tier. The model field in the response will indicate which specific model served the request.
Routing is transparent and fully automatic. You do not need to configure anything — just use the tier aliases and the system handles the rest.
The Starter plan includes access to the Base tier ("fast" alias). This is suitable for teams getting started with AI automation who need reliable, low-latency completions for straightforward tasks.
The SMB plan includes access to both the Base and Mid tiers ("fast" and "standard" aliases). Most teams find the Mid tier covers the vast majority of their automation needs.
The Agency plan includes access to all three tiers: Base, Mid, and Top ("fast", "standard", and "premium" aliases). This plan is designed for agencies and power users who need access to the most capable models for complex client work and high-value workflows.
If you send a request to a tier that your plan does not include, the API will return a 403 Forbidden error with a message indicating the required plan. You can upgrade at any time from the Dashboard.