# Anyone else finding Astra answers too brief and underdeveloped compared with Gemini? — retained record

> Retained reddit record from reddit:ChatGPT.

Canonical: https://highsignal.app/data/records/e0e8d5157722fd28

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:ChatGPT)  published 2026-09-19 13:04 ingested 2026-09-20 02:36

[original source ↗](https://www.reddit.com/r/ChatGPT/comments/1wklqqi/anyone_else_finding_astra_answers_too_brief_and/)

## retained text



```
I’ve been comparing Astra and Astra Light with Gemini using the same questions, and I keep noticing a difference in how much they actually explore the topic.

For example:

Movie recommendations: ChatGPT gives me a short selection with brief explanations. Gemini gives me more to explore and more detail about why each movie might interest me.

History: ChatGPT gives a few short examples, while Gemini develops the stories—the background, what happened, and why it matters.

Follow-up questions: I keep having to push ChatGPT for more depth. Gemini seems to provide that depth without as much prompting.

I’m not saying longer automatically means better, or that Gemini’s facts are always correct. My issue is that ChatGPT’s answers often feel underdeveloped. I want interesting details, useful connections, and explanations that help me learn not just a quick summary.

When I pointed this out, ChatGPT acknowledged that its answers were too brief, but I’m trying to understand why this keeps happening across different topics.

Is anyone else experiencing this with Astra or Astra Light? Is there a setting or custom instruction that actually helps consistently? Or are you finding this is simply a difference in their default response styles?

I’d be interested in other people’s same-prompt comparisons. I want to figure out whether this is specific to my setup or a broader pattern.
```



document fetched 2026-09-20 00:21 · canonical https://reddit.com/r/ChatGPT/comments/1wklqqi/anyone_else_finding_astra_answers_too_brief_and
