Token budget in deepsearch seems to be ignored

[samducker]

When trying to run fast, low-reasoning, low-effort deepsearches, it seems like the token budget setting is being ignored. I’ve tried values like 50,000, 20,000, and 80,000, but the responses still take over 60 seconds—even for fairly simple questions.

For context, ChatGPT is able to answer these same queries within a few seconds.

Is there something I might be missing in the configuration, or is this a known issue?

Any advice would be appreciated—thanks!