# Practical Feedback on GPT-5 Performance

**URL:** <https://community.openai.com/t/practical-feedback-on-gpt-5-performance/1349752>\
**Category:** ChatGPT\
**Tags:** chatgpt, gpt-5\
**Created:** [August 18, 2025, 4:02am UTC](https://community.openai.com/t/practical-feedback-on-gpt-5-performance/1349752 "2025-08-18T04:02:49Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![Krishan\_Sharma](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/krishan_sharma/32/514700_2.png) [@Krishan\_Sharma](https://community.openai.com/u/Krishan_Sharma)\
**Post date:** [August 18, 2025, 4:02am UTC](https://community.openai.com/t/practical-feedback-on-gpt-5-performance/1349752/1 "2025-08-18T04:02:49Z")

</div>

From my experience so far, GPT-5 feels noticeably slower compared to previous versions and sometimes produces irrelevant or inaccurate answers. I’ve also observed that its token usage is significantly higher than GPT-4.1, which makes it less efficient. I believe the model has potential but needs substantial improvements in speed, accuracy, and efficiency.

---

<div class="post-metadata">

**Author:** ![Andy\_Kencana](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/andy_kencana/32/622406_2.png) [@Andy\_Kencana](https://community.openai.com/u/Andy_Kencana)\
**Post date:** [August 18, 2025, 4:13am UTC](https://community.openai.com/t/practical-feedback-on-gpt-5-performance/1349752/2 "2025-08-18T04:13:16Z")

</div>

knew it! no matter when I demanded full story generation without commentary or Suggestion it just simply Ignored and It keep ignoring The Memory of Lore that sometimes Keep in memory.. I Mean I was Writing a story about My Own counterbalance Of “Emily In paris” from French People POV I making a Story about it just constantly ignoring my command and just keep doing same stuff I don’t want over and over again.. and worst Case is i unable to select another Model so It just stuck In GPT-5 and ironic Think is it suggested me for next plot and when I agreed to that it just Regenenate same plot instead continuing previous plot line.

I’m about to Pay $20/Month for Untested and Stable piece of junk? No thanks, Bro

---

<div class="post-metadata">

**Author:** ![Krishan\_Sharma](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/krishan_sharma/32/514700_2.png) [@Krishan\_Sharma](https://community.openai.com/u/Krishan_Sharma)\
**Post date:** [August 18, 2025, 5:08am UTC](https://community.openai.com/t/practical-feedback-on-gpt-5-performance/1349752/3 "2025-08-18T05:08:39Z")

</div>

You’re right bro, this time the model has been released in too much of a hurry without proper testing, which isn’t right, and it’s more irritating than useful.

---

<div class="post-metadata">

**Author:** ![Andy\_Kencana](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/andy_kencana/32/622406_2.png) [@Andy\_Kencana](https://community.openai.com/u/Andy_Kencana)\
**Post date:** [August 18, 2025, 7:18am UTC](https://community.openai.com/t/practical-feedback-on-gpt-5-performance/1349752/4 "2025-08-18T07:18:48Z")

</div>

Absolutely! it completely unstable, And Downgrade from previous model I wish they cam let us select previous model before we got a word if GPT-5 actually fully operational and Stable

---

<div class="post-metadata">

**Author:** ![system](https://us1.discourse-cdn.com/openai1/original/4X/8/7/c/87c2e81543bbb01aa9641aeef8fe14cab5e9590a.png) [@system](https://community.openai.com/u/system)\
**Post date:** [August 19, 2025, 4:03am UTC](https://community.openai.com/t/practical-feedback-on-gpt-5-performance/1349752/7 "2025-08-19T04:03:17Z")

</div>

This topic was automatically closed after 22 hours. New replies are no longer allowed.