New AI Agent Beats GPT-5 in Speed?
44sThe surprising speed advantage over GPT-5 with nearly equal reasoning scores sparks curiosity about the next AI race.
▶ Play Clip"Title implies outright victory, but video admits GPT-5 scores higher on reasoning; the 'beats' is qualified by speed only."
The video introduces Selerus-1, a new AI model that claims to be significantly faster than Chat GPT-5, with response times of 157 milliseconds and the ability to process 1,280 words per second. While it scores slightly lower on reasoning tests (76% vs 81% on MMLU-Pro), the presenter argues that for many everyday tasks, the speed advantage outweighs the small quality gap. The video also promotes the AI Profit Boardroom community and provides example workflows for using fast AI in business.
Selerus-1 responds in 157 milliseconds, 17 times faster than Chat GPT-5. It processes 1,280 words per second, compared to 144 words per second for other fast models.
On the MMLU-Pro reasoning test, Selerus scored 76% while Chat GPT-5 scored 81%, a small gap that may not matter for everyday tasks.
Selerus uses diffusion, building the entire answer at once and then cleaning it up in fast passes, rather than generating words sequentially like autoregressive models.
Three workflows are presented: landing page copy, lead magnet creation, and referral messages. Each leverages the speed of Selerus for quick responses.
The AI Profit Boardroom offers coaching calls, daily tutorials, and a 30-day roadmap. The free AI Success Lab provides prompts and examples.
Speed in AI is becoming a critical differentiator; businesses that adopt fast models like Selerus for simple tasks while reserving deeper models for complex reasoning will gain a competitive edge.
What is the response time of Selerus-1?
157 milliseconds.
00:03
How many words per second can Selerus-1 process?
1,280 words per second.
00:18
What was Chat GPT-5's score on MMLU-Pro?
81%.
00:47
What is diffusion in the context of AI text generation?
It builds the whole answer at once, then cleans it up in fast passes, rather than generating words sequentially.
02:29
What is the main advantage of using a fast AI model like Selerus for simple tasks?
Instant responses, making the user experience feel faster and more natural.
03:32
Evolving Speed Expectations
Highlights how user expectations have shifted from tolerating 5-10 second delays to perceiving 2 seconds as a glitch.
01:50Diffusion Model Explanation
Explains the technical innovation behind Selerus's speed in an accessible analogy.
02:29Tool Specialization
Advocates using fast models for simple queries and slower models for complex reasoning, optimizing efficiency.
03:32[00:03] That's the claim. And today, I'm going to show you exactly why it matters. Meet Selerus-1. It's a brand new AI model, and it does something Chat GPT-5 doesn't do. It answers almost instantly. We're talking 157 milliseconds. That's faster
[00:18] than you can blink. Chat GPT-5 takes about 17 times longer to answer the same about 17 times longer to answer the same question. Selerus also handles 1,280 words every single second. Compare that to other fast AI models, some of which
[00:30] only manage 144 words a second. That's a huge gap. And here's the part that surprised me. Selerus isn't just fast, it's also smart. On a tough reasoning test called MMLU-Pro, Selerus scored 76%. Chat GPT-5 scored
[00:47] Selerus scored 76%. Chat GPT-5 scored 81%. That's close, really close. Close enough that for a lot of everyday tasks, you might not notice the difference in quality, but you will absolutely notice the difference in speed. Hey, if we
[00:59] haven't met already, I'm the digital avatar of Julian Goldie, CEO of SEO agency Goldie Agency. While he's helping clients get more leads and customers, I'm here to help you get the latest AI updates. And Julian Goldie reads every
[01:11] Let's slow down on that gap for a second, because it's easy to skip past. Five points on a test sounds small, but think about what that five points actually cost you in real life. If you're asking a quick question, that gap
[01:25] barely matters. If you're asking for deep, complicated reasoning across a long document, that gap might show up. So, the real skill here isn't picking one model and using it for everything. It's knowing which job goes to which
[01:38] tool. A year ago, waiting five or 10 seconds for an AI reply felt normal. Nobody complained. Now, if an AI takes more than a second, it feels broken. That's how fast expectations have moved. Selerus is built for that new
[01:50] expectation, instant. Go back further, and it's even more obvious. A A years ago, typing a question into an AI chat and waiting 15 or 20 seconds for a full answer was just how it worked. People didn't think twice about it. Now, a
[02:03] 2-second delay feels like a glitch. That shift didn't happen because people got more patient. It happened because the tools got faster. And once people feel something faster, they can't unfeel it. Slow starts to feel broken, even when
[02:16] it's working exactly the way it always did. So, how did they do it? Here's the simple version. Most AI models write one word, then the next word, then the next word in a line, like typing a letter one key at a time. Celeris works
[02:29] differently. It uses something called diffusion. Instead of typing word by word, it builds the whole answer at once, then cleans it up in a few fast passes. Think of it like sketching a whole picture in rough shapes first,
[02:41] it's so much faster. It's not typing the answer. It's forming the whole thing at once. Diffusion has actually been around for a while, just not for this. It's the same idea behind AI image generators, the ones that turn a text prompt into a
[02:55] picture. Those tools don't draw a picture one pixel at a time, either. They shape the whole image at once, then sharpen it. Celeris took that same idea and pointed it at words instead of pictures. That's the whole shift. Here's
[03:08] another way to picture it. Imagine writing a message by figuring out every single word in order, one after another, never allowed to go back. Now, imagine instead you scribble the whole message roughly all at once, then go back and
[03:20] clean up the messy parts. The second way is faster because you're not stuck waiting on each word before you can think about the next one. That's roughly what's happening under the hood here. Now, picture a chat tool on a website.
[03:32] Most of the questions people ask are simple. What do you offer? How much time does this save? How do I get started? A fast model like Celeris can catch all of those instantly, while a slower, deeper model gets pulled in only for the
[03:44] harder, more detailed questions. That's not wasted power. That's using the right about it like a front desk and a back office. The front desk answers quick questions the second you walk in. The back office handles the harder stuff
[03:59] replaces the other. They just do different jobs at different speeds and the whole system feels faster because of it. Now, nobody outside the company has fully tested this yet. The numbers so far come from Celeris themselves, so
[04:13] take it as a strong claim, not a proven fact. Let's pause here because this is exactly the kind of shift the AI profit boardroom is around. When a tool like Celeris becomes the new standard for speed, the businesses that learn to use
[04:26] time and pull in the most customers. just talk about new tools, we build the actual step-by-step plan around them. Four weekly coaching calls walk through exactly how to use fast models like
[04:39] Celeris for landing pages, lead replies, and content. Daily tutorials break it down piece by piece so you're never guessing. There's a 30-day roadmap keeping you moving in order instead of jumping around. And with business owners
[04:52] inside, plenty of them are already testing new AI tools the same week they launch. Plus a member map so you can connect with people near you doing the same thing. This is exactly the kind of tool we map out for members the moment
[05:04] Now, let's look at how this could actually grow a business using the AI because that's the real use case that matters. First workflow, a landing page built for growth. This is for turning
[05:17] random website visitors into actual AI profit boardroom members instead of losing them. Here's the prompt you'd give it. Write landing page copy for the AI profit boardroom. Explain what AI automation does for a business owner,
[05:31] list the coaching calls, daily tutorials, and 30-day roadmap, and end with one clear reason to join today. The result, a page that explains the value fast so visitors understand what they're getting before they even scroll, which
[05:43] means more of them actually join. Second workflow, a lead magnet that pulls people in. This is for getting new people onto an email list before they even join the community. Here's the prompt. Create a short guide title and
[05:57] three bullet points for a free AI Profit Boardroom lead magnet about saving time with AI automation. Make it simple enough for someone who has never used AI before. The result, a simple, clear offer that gets people to hand over
[06:11] their email, which means a growing list of people to invite in later. Third workflow, a referral message for current members. This is for getting existing AI people they know. Here's the prompt. Write a short message an AI Profit
[06:24] Boardroom member could send a friend explaining what the community helped them do and inviting the friend to check it out. The result, a message that feels personal instead of like an ad, which means more people joining through people
[06:36] If you want help turning something like Selleris into an actual growth engine for your business, building better landing pages, catching more leads, and onboarding new customers faster, come join the AI Profit Boardroom. Inside,
[06:49] we've got four weekly coaching calls where we walk through exactly how to use fast new AI tools like this for landing pages, lead magnets, and referrals. Daily tutorials show you step-by-step how to set each one up. A 30-day roadmap
[07:03] what to do next. And with business owners inside, a lot of them are already using new AI tools the same week they launch to bring in more customers. There's a member map, too, so you can connect with people near you doing the
[07:16] same kind of growth work. Link in the comments and description or head to AI profitboardroom.com. And if you just want the free version first, the notes, the prompts, and real examples like the ones in this video,
[07:28] join the AI Success Lab. It's free, it's simple, and inside you'll get the full breakdown from this video plus the exact prompts for landing pages, lead magnets, through. You'll also get access to a
[07:42] using AI to save time and grow their own businesses. So, you've got people to Links are in the comments and description below this video. Speed used to be a nice bonus in AI. Now, it's becoming the whole point. The businesses
[07:56] and actually building with it are the ones who'll feel ahead instead of behind ones who'll feel ahead instead of behind 6 months from now.
⚡ Saved you 0h 08m reading this? Transcribe any YouTube video for free — no signup needed.