Nvidia/llama-3.1-nemotron-nano-4b-v1.1 Tool Calling Issue

Nvidia/llama-3.1-nemotron-nano-4b-v1.1 Tool Calling Issue

I am using the nvidia/llama-3.1-nemotron-nano-4b-v1.1 model within an n8n AI Agent workflow. The model is not correctly executing the tools connected to the agent. Instead of triggering the tool, it outputs the tool call information as a plain text string directly in the chat response.

I have attempted to use this model in two different ways, both with issues:

Using NVIDIA NIM via the OpenAI Node:

When I connect to the NVIDIA NIM’s OpenAI-compatible endpoint, the model correctly identifies which tool to use but fails to execute it. It returns a string like “OOLCALL>[{“name”: “Database_FastAPI”, “arguments”: {“input”: “your-policy-number-here”}}]</TOOLCALL” as a message instead of triggering the actual tool node. This is the primary issue, as shown in the attached screenshots.

Using Ollama:

I pulled a community version of the model (avil/nvidia-llama-3.1-nemotron-nano-4b-v1.1:thinking-off) and used the Ollama Chat node. This method has partial success:

  1. It can successfully call a single RAG tool when it’s the only tool.
  2. However, when multiple tools are connected (e.g., RAG + a custom tool), it exhibits the same failure as the NIM method, outputting the tool call as a string.
  3. Additionally, despite the model version name (thinking-off) and explicit instructions in the prompt to be concise, it still outputs “thinking” or “reasoning” steps before its final answer.

The expected behavior is for the n8n agent to recognize the model’s intent to use a tool, parse the tool name and arguments, and execute the corresponding tool in the workflow.

My workflow with NIM credentials in openai node:

It’s Output:


My Workflow with ollama community model is the same as I posted above. (I have used OpenAI node for this too ):

It’s Output:

Single tool: Tool call success. But thinking mode on despite setting it off in the prompt and defining the appropriate payloads.


Multiple tools: Tool call failure

Information on your n8n setup

  • n8n version: 1.93.0
  • Database (default: SQLite): SQLite
  • n8n EXECUTIONS_PROCESS setting (default: own, main): Own
  • Running n8n via: npm (exposed it with ngrok)
  • Operating system: Linux ubuntu
1 Like