The Launch Failure of GPT-5 Shows OpenAI is in Decline

After such a long wait, we get one of the strangest Model launches in history, and OpenAI is reeling from the fall-out 😲

Screenshot of GPT-5 Live event. August 7th, 2025.

Hey Everyone,

Welcome to another article from AI Supremacy, a Newsletter about AI at the intersection of tech, business, society and the future. Welcome to our new readers who are joining 170,000 others. Check the rising publications on our channel of Technology here. Welcome to join our new community.

• (Eds: Updated August 11th, 2025) I’ve updated this article with a lot more details, links and ongoing debates about GPT-5 for clarity.

This article is intended as an introduction to some of the issues OpenAI with the launch of GPT-5 are facing from many varied voices, including OpenAI’s own PR. How the two contrast is fairly striking in August, 2025.

Should Sam Altman be Worried? 🤦‍♂️😞

Showcases

Full disclosure, community posts since the GPT-5 launch around 5 days ago have tended towards more negative about GPT-5 than positive):

1. GPT-5 is no slowdown (a lot of interesting context and links here)

2. GPT-5 And Privacy by Design: Does OpenAI Care?

3. How ChatGPT accidentally became the fastest-growing product in history

4. GPT-5: Overdue, overhyped and underwhelming. And that’s not the worst of it.

5. GPT-5 at Scale: Why Reliability Slipped

6. GPT-5 Disappoints

7. The Complete ChatGPT-5 Review: 5 Real-World Tests and the Playbook to Use It Right

8. OpenAI’s Open-Weight Models: Overhyped & Shockingly Underperforming

So I listened to OpenAI’s 1 hour live stream about GPT-5’s launch, watch it here. Here is the 3min recap (edited by C.Grennan).

GPT-5 Launch 3-min Recap

Some of GPT-5’s talking points are crazy exaggerated, but that’s the world we live in now, so please take some of these quotes with a grain of salt. This article is fairly long, so click on the title or click below to read it on the web (it will be easier to see the infographics).

(Like) “Having a team of PhDs in your pocket. “

• As of the second week of August, 2025 - GPT-5 is rolling out today to Free, Plus, Pro, and Team users, with Enterprise and EDU next week.

• It tooks OpenAI 162 days to get from GPT-4.5 to GPT-5. GPT‑4.5 was officially released on February 27, 2025. But the wait for GPT-5 felt like years! That’s also because Anthropic, Google and Qwen caught up, and we haven’t even gotten DeepSeek-R2 yet.

• It took OpenAI, GPT-4 (March 14, 2023) to GPT-5 (August 7, 2025) a grand total of 877 days to get from GPT-4 to GPT-5.

The “Benchmarkmaxxing”Crisis

As pointed out by

Gennaro Cuofano

via Andrej Karpathy, GPT-5 so far has been peak Benchmarkmaxxing whereby agentic workflows are prioritized over customer experience. Even the actual usage of reasoning models in the previous iteration of OpenAI’s models were shockingly low.

TL;DR

OpenAI’s GPT-5 may not be as good as we had hoped, but it’s a product suited for their unique approach to AI.

1. Unified System with Auto-Switching

2. Advanced Reasoning Capabilities

3. Enhanced Multimodal Functionality

4. Improved Coding Performance

5. Reduced Hallucinations and Improved Safety

6. Customizable Personalities and Interface

7. Expanded Context Window: 256k

8. Integration with External Tools

9. Sycophancy Reduction

10. Accessibility and Pricing: API pricing is now at $1.25/million input tokens and $10/million output tokens, with reasoning tokens counted as output.

“It is a unified system that automatically switches between providing a quick response and taking time to reason through a hard problem to provide the best answer.” - Srinivas Narayanan

• It appears as if OpenAI prioritized capacity saving over its customer experience as the auto-switching routing-experience has been a disaster:

Some Controversy 💥

“Suddenly retiring every other model without warning was a weird move by OpenAI.

… and they did it without explaining how switching models worked or even details of various GPT-5 model and what older models they replaced.

…and they did it after many built workflows & training & assignments around older models, maybe breaking them. Do my old GPTs work? Who knows!” - Ethan Mollick

Top AI book about OpenAI

In Empire of AI, Karen Hao exposes OpenAI’s transformation from an altruistic nonprofit to a profit-driven entity, revealing its secretive culture, exploitation of global labor, and significant environmental costs. The book critiques the AI industry’s imperialistic tendencies, drawing parallels to historical colonialism while highlighting internal power struggles and ethical dilemmas under Sam Altman’s leadership.

Quote of the Day

“Let’s take advantage of this moment of clarity, and forget the speculative futures. If we wake up to millions of addicted and deluded AI chatbot users, students incapable of finishing their homework without help from an app, and automation software that surveils and immiserates workers, each hurriedly installed on the top layer of our society, well, the joke will have been on all of us.” -

Brian Merchant

Mostly a Quality of Life upgrade 💫

GPT-5 is not a breakthrough, but an iterative improvement over GPT-4.5, with some areas, like writing, reportedly weaker than predecessors. While GPT-5 is a significant step forward, it’s not a revolutionary leap toward artificial general intelligence (AGI) or a meaningful improvement outside of some quality of life features. That’s not to say that the slick interface and improved customization in design, voice and personality isn’t significant.

OpenAI’s ChatGPT Personas 😂

• Preset Personalities: Users can choose from four new preset personalities to customize how ChatGPT interacts:

  • Cynic (sarcastic, blunt)

  • Robot (dry, precise)

  • Listener (calm, supportive)

  • Nerd (curious, explanatory)
    These are opt-in, adjustable anytime, and create different conversational styles without needing custom prompts.

Quick Take - Reasonable API Cost

• $1.25 input / $10 output per million tokens - See Full pricing list.

• Some are claiming that the main differentiator of GPT-5 is on price. Is price enough?

So much has already been written about GPT-5 in a short time since its launch, I’m hoping this articles give you a broad selection of potential sources to browse from:

Showcases Continued

1. GPT-5 is a joke. Will it matter?

2. GPT-5s Are Alive: Basic Facts, Benchmarks and the Model Card

3. GPT-5 Hands-On: Welcome to the Stone Age

4. GPT-5: It Just Does Stuff

5. GPT-5 and the arc of progress

6. GPT-5: Key characteristics, pricing and model card

7. The surprise deprecation of GPT-4o for ChatGPT consumers

8. 🔮 Feel the AGI yet?

9. An exclusive inside look at GPT-5 - Listen on YouTube.

The quotes in OpenAI’s PR are fairly jarring:

“It’s beyond a collaborator, it’s almost like a mentor.” (see video below of Medical Researcher)

OpenAI created tons of Videos and Blogs about GPT-5

Some of these are fairly speculative and subjective:

1. GPT-5 and the new era of work

2. Introducing GPT‑5 for developers

3. GPT-5 for Scientific Discovery (video)

4. Empowering a Medical Researcher (video)

“Fundamentally it will change the way we do Science”.

• But in reality?

Great Customization and Personalization

I think GPT-5’s memorable part is not its unified system but rather the moves in the personalization and customization direction. Easily Voice Mode to me gets the highest marks here:

• Voice Mode Customization: The updated Voice mode can adjust tone, pace, and response length based on user instructions. Voice features are more adaptive and available to all users with higher usage limits for paid subscribers. Voice supports custom GPTs but currently retains a standard default voice behavior separate from the personalities available in text.

• One word answer feature is super engaging for me!

• OpenAI seems to have noticed that Health questions are a big deal for ChatGPT users.

Share your Take:

Share your insights on this note:

“It’s been a great year for health AI, both for performance at the frontier and at cost. We’ve gone from 0% (GPT-4o) to 46% (GPT-5 thinking) on HealthBench Hard, a health benchmark built with 250+ doctors.” - Karan Singhal

Anatomy of a GPT-5 Prompt

Credit: Ben Hylak

GPT-5 Top Use Cases

1. Voice mode is SOTA

2. Can be used for vibe coding

Who is BigAI? ✨

BigAI are the new companies most likely to challenge BigTech in how they leverage Generative AI. Based on my recent observation of the AI ecosystem these are the likely challengers in 2026:

1. OpenAI

2. Google DeepMind

3. Anthropic

4. xAI

5. Meta Superintelligence Lab

6. Thinking Machines Lab

7. DeepSeek

8. Qwen

BigAI are for me the leading labs who will likely build the best consumer and Enterprise AI products around AI. I do not consider Apple, Amazon, Microsoft to be in this group. BigAI are likely disruptive for other BigTech companies who aren’t well diversified like potentially Meta. Microsoft and Amazon are fairly well diversified companies.

Why Ecosystem Competitiveness matters?

It’s somewhat unlikely OpenAI will be able to remain competitive or a frontier lab given AI talent poaching and competitive pressures moves forward as Anthropic plans to speed up iteration of model development and Google and Qwen have hit new cadences of LLM iteration difficult for OpenAI to match. As far as I can tell, 2026 will represent competitive pressures on OpenAI is has not witnessed since it launched ChatGPT in late 2022. The general capabilities of GPT-5 is very worrisome for OpenAI’s ability to remain relevant as a frontier lab.

The public criticism of GPT-5 remains fresh in our psyche during the first week after the launch.

What does Reddit say about GPT-5? 🤔

• Read some comments on Reddit: GPT-5 is a disaster.

• I’m sorry but I’m being reasonable

• GPT-5 is a massive downgrade.

• GPT-5 will be better in a lot of fields.

• GPT-5 thoughts? (later in the day these types of comments began to appear on X).

🔴 YouTube GPT-5 Hot takes

• OpenAI’s own YouTube Short promo. (OAI)

• Surprising developers with GPT-5. (OAI)

• Matt Wolfe

• Matthew Berman

• Wes Roth

• Powering Creativity (writing) with GPT-5 (OAI)

• Sam Altman interview with Cleo Abram. (about GPT-5)

OpenAI GPT-5 Feedback Is Clear

So much great feedback about GPT-5, but I doubt OpenAI or Sam Altman will listen. Many users and devs report GPT-5 does not reach parity with Claude Opus 4.1 in coding capabilities. The scrutiny GPT-5 is facing is an evolving story.

GPT-5 Lagged Grok 4 in ARC-AGI-2

Grok 4 is still state-of-the-art on ARC-AGI-2 among frontier models, according to

François Chollet

.

15.9% for Grok 4 vs 9.9% for GPT-5.

These were among the (many) signs that GPT-5 was really not SOTA and OpenAI has slipped considerably in its capabilities in the 2024-2025 mid period.

Francois Chollet - https://x.com/fchollet/status/1953511631054680085

If OpenAI’s best models are not SOTA or frontier any longer, what does it mean for the future of ChatGPT and their agentic systems? 🤔 OpenAI CEO, Sam Altman, told CNBC last week that artificial general intelligence is "not a super useful term." I wonder why he’s be saying that?