Runtype
Settings

Available AI models

Runtype gives you access to hundreds of AI models from leading providers, ready to use in your Flows, Agents, and Products. Many popular models work instantly with platform keys, and you can unlock the full catalog by connecting your own provider keys.

New accounts include a curated set of high-quality models so you can start building right away. To see what is available for your workspace, go to Settings → Models. For an overview of the settings area, see What’s in Settings?.

Platform keys vs. Bring Your Own Key (BYOK)

Platform keys are included with every Runtype plan. They give you immediate access to popular models from providers like OpenAI, Anthropic, Google, and xAI without adding your own API keys. Usage is tracked through Runtype, and you can review account usage in Usage data.

BYOK (Bring Your Own Key) lets you connect your own provider API keys for full access to supported models. This is useful if you already have provider agreements, need access to specialized models, or want billing to happen through the provider. BYOK is available on Growth plans and above. For plan details, see Billing and plans.

To connect your own keys, go to Settings → Models and open the My Providers tab. For setup steps, see Connecting AI model providers.

Supported providers

Runtype supports models from a wide range of providers, and the catalog is updated regularly as new models are released.

  • Included with platform keys: OpenAI, Anthropic, Google, xAI, Vercel, and Meta via Mixlayer.

  • Available with BYOK: DeepSeek, Mistral AI, Together AI, Qwen, Perplexity, AWS Bedrock, Google Vertex AI, Zhipu AI, Tinfoil, and other supported providers.

You can browse the available models in Settings → Models. Use search and filters to find models by name, provider, or capability.

Choosing the right model

  • General-purpose tasks: Claude Sonnet 4.5, GPT-5, and Gemini 2.5 Pro balance quality and speed.

  • Fast and cost-efficient: GPT-5 Mini, Claude Haiku 4.5, Gemini 3 Flash, and Grok 4 Fast work well for high-volume or latency-sensitive Flows.

  • Complex reasoning: o3, DeepSeek R1, and Qwen 3 Max Thinking are strong choices for multi-step reasoning.

  • Code generation: GPT-5.1 Codex, Devstral, and Qwen 3 Coder are built for writing and analyzing code.

  • Vision and images: GPT-5, Claude Opus 4.6, Gemini 2.5 Flash, and Grok 2 Vision can process images alongside text.

If you are new to Runtype, start with a simple Flow and test different models against the same prompt so you can compare speed, quality, and cost.

Smart model routing

When you use a model name like claude-sonnet-4-5 or gpt-5 in a Flow, Runtype routes the request to the best available provider for that model. You do not need to use provider-specific model IDs for common cases.

Next steps

Was this helpful?