n8n AI Agent Node v2: native support for Claude 4, Gemini and DeepSeek R2

Core conclusion

On May 18, 2026, n8n released a major update to AI Agent Node v2, with native access to the three major model APIs of Claude 4, Gemini 2.5 Pro and DeepSeek R2. This is n8n’s most significant upgrade to AI nodes to date—the free community version can use all Agent functions, and users can build an automated pipeline for multi-Agent collaboration without writing a line of code.

Key Points

• Time of incident: 2026-05-18
• Scope of influence: All n8n users (self-built/cloud/community version)
• Core changes: AI Agent Node is upgraded to support Claude 4, Gemini 2.5 Pro, DeepSeek R2, and a new visual multi-Agent workflow editor is added.

Background and trigger events

n8n is one of the most popular open source automation platforms out there, with a large user base among AI content creators and independent developers. Previously, n8n's AI Agent Node mainly supported the OpenAI series models. Access to other mainstream models required additional HTTP nodes or third-party packaging modules, which increased the threshold for building complex workflows.

There are three core driving factors for this v2 upgrade: First, the significant improvement in reasoning capabilities and context length of Claude 4 and Gemini 2.5 Pro has sharply increased n8n users' demand for multi-model selection; second, after the release of DeepSeek R2 open source weight, the demand for low-cost local model access from the deployment group has increased significantly; third, the demand for multi-Agent collaboration scenarios in the n8n community is increasingly strong.

Key Impact (by Dimension)

Adaptation suggestions

In response to this upgrade, the content operations team can do three things immediately:

• Reconstruct existing workflow: Split the original pipeline that relied on a single model into a multi-Agent collaboration model - DeepSeek does structured tasks, Claude does high-quality output, and Gemini handles multi-modal input
• Use DeepSeek R2 to control costs: For collection/classification tasks with thousands of calls per day, API costs can be reduced by 80-90% after switching to DeepSeek R2
• Build Agent Quality Gate Pipeline: Use Claude 4 as content review Agent to perform secondary verification on automatically generated content, forming a closed loop of "generate → review → release"

Implementation Task List

• Upgrade n8n to the latest version (v1.75+ including AI Agent Node v2)
• Configure the API Key of Claude 4 / Gemini 2.5 Pro / DeepSeek R2 in Credentials
• Create a test workflow containing multiple Agents to verify the model switching logic
• Add routing logic to the production workflow: automatically select the optimal model based on task type

Tool entry

n8n, OpenAI, ChatGPT, Claude, Gemini, DeepSeek, Ollama, LM Studio

Internal link guidance

• Want to learn how? First look at the n8n MCP build guide from scratch: AI Agent-Driven Content Automation: n8n MCP Building Guide from Scratch
• Existing workflow upgrade reference: How to build an AI content automated distribution system with n8n + ChatGPT: a complete 30-minute tutorial
• Real case: n8n + OpenAI content automation website monthly income $4,500 → I used n8n + OpenAI to build AI content automation website: a complete review of monthly income $4,500
• Advanced gameplay: Claude + n8n achieve monthly income in half a year $12,000 → He Built an AI Automation Stack with Claude + n8n — $4K to $12K/mo in 6 Months