Significant Token Stream Lag on 5.1 Thinking and 5.2 Thinking Models

[elfin falcon]

Reported by @formal aspen

Bug Report: Significant Token Stream Lag on 5.1 Thinking and 5.2 Thinking Models

`Steps to Reproduce`

1. Select 5.1 Thinking or 5.2 Thinking
2. Enter a prompt, like 'Testing'
3. Await output response.

Context: Prompts regardless of complexity output with significant lag. The model will report the response takes ~30 seconds to generate but minutes at time to output, even for simple requests, where the token stream is spitting out words at a crawling pace. This occurs on both GPT Web and iOS on my account for the 5.2 Thinking and 5.1 Thinking Families.

`Expected Result`

Normally, one would expect Thinking products to take seconds to minutes to think, but the output stream responds over the course of several seconds to minutes, but the stream itself is contiguous, not lagging and printing words at a 'drip' pace.

`Actual Result`

Prompts regardless of complexity output with significant lag. The model will report the response takes ~30 seconds to generate but minutes at time to output, even for simple requests, where the token stream is spitting out words at a crawling pace. This occurs on both GPT Web and iOS on my account for the 5.2 Thinking and 5.1 Thinking Families. No other products (5.x Instant, 4.1, 4o, appear to be affected on my account).

`Environment`

Chat GPT for iOS (1.2026.013), Chrome (144.0.7559.11)

Jan 31, 2026, 1:01 PM

Additional Information

Please provide relevant details to help resolve the issue, such as:

• ChatGPT Shared Link (if applicable).
• Screenshots or videos demonstrating the problem.

-# ➜ Need to contact support? Visit the OpenAI Help Center.