How do one know which model supports parallel tool-use (i.e., calling multiple tools at the same time)?
From my experiments, GPT-4o supports this and O3 doesn’t, is there any documentation for parallel tool-use support?
1 Like
I think the promise is that they all might do it unless you disable it.
https://platform.openai.com/docs/guides/function-calling?api-mode=responses
Parallel function calling
The model may choose to call multiple functions in a single turn. You can prevent this by setting parallel_tool_calls to false, which ensures exactly zero or one tool is called.
Note: Currently, if you are using a fine tuned model and the model calls multiple functions in one turn then strict mode will be disabled for those calls.
Note for gpt-4.1-nano-2025-04-14: This snapshot of gpt-4.1-nano can sometimes include multiple tools calls for the same tool if parallel tool calls are enabled. It is recommended to disable this feature when using this nano snapshot.
1 Like
Parallel tool calling is a wrapper - an additional tool that is placed called multi_tool_use, within which function names of the developer can be placed.
Note the difference:
- tool: OpenAI internal tools
- function: a list of developer functions all placed in one type of tool called “functions”
The primary way you can see if it is working, to see if the additional tool is being placed into AI context and is language offered to the AI (which often cannot understand or properly use the tool description that comes along with multi_tool_use) is to look at the input token usage billed, first with the difference of manually setting the parallel tool parameter for the endpoint true or false, then with switching the AI model - see if the input is increased and reduced or errored or silently fails to deliver..
The unaddressed concern previously demonstrated:
1 Like