Is GPT-5.1 broken or is it just me? Web browsing never triggers + ignoring instructions

Understanding Frustrations with GPT-5.1: Persistent Browsing and Instruction Issues

In recent months, many users leveraging GPT-5.1 have reported ongoing challenges that impact their experience and productivity. A common concern revolves around the model’s handling—or rather, mishandling—of browsing capabilities and instruction adherence. Here’s a detailed overview of these issues, their implications, and potential avenues for resolution.

Key Challenges Reported by Users:

1. Browsing Functionality Malfunction:
2. When prompts such as “search online,” “check the internet,” or similar phrases are issued, the model responds affirmatively, indicating it will look up information.
3. However, no actual browsing occurs; the browsing tool does not activate.
4. Consequently, no search results are retrieved, rendering the feature ineffective.

5. These behavior patterns persist across varied conversations and sessions, suggesting a systemic issue.

6. Ignored Instructions and Workflow Interruption:

7. Users have observed that the model sometimes disregards straightforward, explicit instructions.
8. This inconsistency can cause disruptions in workflows, especially when specific responses or actions are expected.
9. Activating browsing capabilities appears unreliable unless users employ unusually phrased prompts or reset conversations, indicating potential underlying regressions.

Platform and Version Context:

• The problems are notably present in the Windows desktop application for GPT-5.1.
• The issues have been persistent over several months, not isolated incidents or recent updates.

Implications for Users:

These malfunctioning features and instruction inconsistencies can significantly hinder productivity, particularly for professionals relying on real-time information retrieval and accurate adherence to directives. As such, users are concerned about the stability of GPT-5.1 and are seeking solutions or confirmation of widespread bugs.

Seeking Solutions and Community Input:

• Are other users experiencing similar issues with GPT-5.1’s browsing capabilities and instruction compliance?
• Have any fixes or workarounds been shared by the community or official support channels?
• Is this indicative of a regression that might be addressed in upcoming updates?

Conclusion:

Persistent issues with browsing activation and instruction adherence in GPT-5.1 highlight the importance of ongoing software testing and user feedback. For professionals depending on these features, such challenges can be highly disruptive. Staying engaged with official updates and community forums can be beneficial in finding workarounds or awaiting future fixes. If you are encountering similar problems, sharing your experiences can help drive awareness and potentially expedite solutions from the developers.

By maintaining open communication and collaborative troubleshooting,