Made with these models
The 3 models this integration runs.
claude-haiku-4-5 on claude
claude-haiku-4-5text-to-text
claude-opus-4-6 on claude
claude-opus-4-6text-to-text
claude-sonnet-4-6 on claude
claude-sonnet-4-6text-to-text
claude-haiku-4-5 on claude
claude-haiku-4-5text-to-text
claude-opus-4-6 on claude
claude-opus-4-6text-to-text
claude-sonnet-4-6 on claude
claude-sonnet-4-6text-to-text
claude-haiku-4-5 on claude
claude-haiku-4-5text-to-text
claude-opus-4-6 on claude
claude-opus-4-6text-to-text
claude-sonnet-4-6 on claude
claude-sonnet-4-6text-to-text
Best-in-class language model for text generation and reasoning.
- Frontier reasoning, structured output and code review
- Drives downstream image, 3D and delivery nodes
- Charged on Anthropic-reported token usage
- Bring your own Anthropic key, or run on the platform key
Claude is the reasoning workhorse of a Sushibelt pipeline. Drop an LLM Worker on the canvas and Claude will write the prompt for your image model, turn a loose brief into structured JSON, summarise a document, review generated code, or decide which branch a Conditional Switch should take.
It reads anything upstream — text, extracted JSON, a prior model's output — and its own output can drive any downstream node, which is what makes multi-model orchestration work: Claude plans, an image model renders, a 3D model promotes the result, and a delivery node commits it to your repo.
Billed on the tokens Anthropic actually reports for the call, not an estimate. Bring your own Anthropic key per project, or run it on the platform key.
llmreasoningtextstructured-outputclaudeanthropicorchestration