"Error processing request: (invalid_value) 'tool_outputs' too large: the combined tool outputs must be less than 512kb. Code: invalid_value Message: 'tool_outputs' too large: the combined tool outputs must be less than 512kb."

Question

"Error processing request: (invalid_value) 'tool_outputs' too large: the combined tool outputs must be less than 512kb. Code: invalid_value Message: 'tool_outputs' too large: the combined tool outputs must be less than 512kb."

Nishith Reddy Gurram 0

We’re encountering the following error when the tool output is large:
"Error processing request: (invalid_value) 'tool_outputs' too large: the combined tool outputs must be less than 512kb. Code: invalid_value Message: 'tool_outputs' too large: the combined tool outputs must be less than 512kb."

Is this limitation related to the subscription plan?

Is there any configuration option at the agent level to increase this limit?

If not, what’s the recommended approach to handle large tool outputs?

Aryan Parashar 3,695 Reputation points Microsoft External Staff Moderator

2025-12-01T06:10:05.94+00:00

Hi Nishith Reddy Gurram,

Yes, this is a known limitation in both the OpenAI API and the Azure OpenAI API. Both platforms impose a maximum combined tool-output size limit (currently 512 KB).

To avoid this error, the recommended approach is to ensure that your tool returns a constrained, pre-filtered, or paginated response.

In other words, the tool should only return the minimal data required for the model to proceed, and any large payloads (e.g., file contents, large JSON objects, or unbounded lists) should be truncated, summarized, or handled in smaller chunks on your end.

Let me know if you still have further queries.
Thank you for reaching out to The Microsoft Q&A Portal.

1 answer

Your answer

Aryan Parashar 3,695 Reputation points Microsoft External Staff Moderator

2025-12-01T06:10:05.94+00:00

Hi Nishith Reddy Gurram,

Yes, this is a known limitation in both the OpenAI API and the Azure OpenAI API. Both platforms impose a maximum combined tool-output size limit (currently 512 KB).

To avoid this error, the recommended approach is to ensure that your tool returns a constrained, pre-filtered, or paginated response.

In other words, the tool should only return the minimal data required for the model to proceed, and any large payloads (e.g., file contents, large JSON objects, or unbounded lists) should be truncated, summarized, or handled in smaller chunks on your end.

Let me know if you still have further queries.
Thank you for reaching out to The Microsoft Q&A Portal.

Answer 1

This is a hard platform limit, not related to your subscription plan.

Plan-related? No. The 512 KB combined tool_outputs limit is fixed.

Configurable? No. There’s no agent or org setting to increase it.

Recommended approaches:

Chunk or paginate tool outputs and process them incrementally.

Summarize or filter results before returning them to the agent.

Store large results externally (e.g., object storage/DB) and return a reference or ID.

Stream results if supported, instead of returning everything at once.

These patterns avoid hitting the limit while keeping workflows reliable.

Share via

"Error processing request: (invalid_value) 'tool_outputs' too large: the combined tool outputs must be less than 512kb. Code: invalid_value Message: 'tool_outputs' too large: the combined tool outputs must be less than 512kb."

1 answer

Your answer