ChatGPT

The Generic variant provides a direct API connection optimized for standard OpenAI specifications and baseline latency, while the OpenRouter variant provides cost-optimized routing with automated prompt caching discounts and multi-model fallbacks for high-volume systems.

Choose the direct Generic variant when you require standard parameters, predictable direct routing, and baseline gpt-5.1 defaults. Select the OpenRouter variant when running high-throughput production pipelines that benefit from automated prompt caching cost reductions up to 90% and dynamic routing fallbacks.

Use reasoning models like o3 or gpt-5.4 for rigorous multi-step STEM problems, complex code verification, and mathematical proofs where scaled test-time compute is essential. Use gpt-5.1 or gpt-5.2 for balanced conversational tasks, knowledge synthesis, and general-purpose reasoning with faster response times.

Avoid this node for offline, local, or air-gapped on-premise environments because all supported models are cloud-hosted. It is also not suitable for sub-millisecond real-time requirements or using heavy reasoning models on simple single-turn lookups where mini variants are more efficient.

This node accepts text prompts alongside image attachments and document references for multimodal analysis. Multimodal models like gpt-5.4, gpt-5.2, gpt-4o, and o4-mini can inspect high-resolution images, technical drawings, and text files directly.

Set the temperature control close to 0 or down to -1 and provide a static seed integer. Pairing a fixed seed with low temperature minimizes randomness during token sampling, ensuring reproducible code refactors and strict JSON schema adherence.