Version

v1.6.0-beta2-24-g998d535f

Which installation method(s) does this occur on?

Source

Describe the bug.

When using streaming endpoints with ReAct agent (either HTTP or WebSocket), the full response is displayed all at once instead of being streamed out token-by-token. This results in significantly increased perceived latency before the user sees any response, even if TTFT is low.

Relevant log output

[Paste the error here, it will be hidden by default]

Other/Misc.

The Tool-Calling Agent streams tokens correctly under the same configs.

Code of Conduct

• I agree to follow the NeMo Agent Toolkit Code of Conduct
• I have searched the open bugs and have found no duplicates for this bug report