Skip to content

[Bug]: change embedding model causing infinite typing status in Telegram #27143

@paul-tian

Description

@paul-tian

Summary

After changing the default embedding model (text-embedding-3-small) to text-embedding-3-large, openclaw seems stuck at replying.

Steps to reproduce

  1. Change the embedding model in openclaw.json from "text-embedding-3-small" to "text-embedding-3-large"
  2. restart the gateway
  3. send message

Expected behavior

  1. reply the message, then the "typing..." status finish
  2. can respond to the following message

Actual behavior

  1. reply the message
  2. keep at "typing..." status
  3. not responding to the new message
  4. use /stop will lead to a repeat reply
  5. change the embedding model back to text-embedding-3-small and restart the gateway can recover.

OpenClaw version

2026.2.24

Operating system

macOS 26.3

Install method

curl -fsSL https://openclaw.ai/install.sh | bash

Logs, screenshots, and evidence

Impact and severity

No response

Additional information

No response

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't workingstaleMarked as stale due to inactivity

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions