GPT 5.4 First Test Results

0:00Today on the AI Daily Brief, GBT 5.4 is here, and these are both the first impressions from the broader world as well as my first test results. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. Alright, friends, quick announcements before we dive in. First of all, thank you to today's sponsors, KPMG, AIUC, Blitzy, and Prompt QL. To get an ad-free version of the show, go to patreon.com slash AI Daily Brief, or you can subscribe on Apple Podcasts. If you are interested in sponsoring the show, send us a note at sponsors at AIDailyBrief.ai. Lastly, as often happens, when we have a big new model release, no headlines today, we are just going to spend all of our time on this exciting new model. So without any further ado, let's dive in. After a couple of weeks where mostly we've been talking about big macro issues like the Pentagon and Anthropic and all of that sort of thing, we finally have the cool fresh breeze of a new exciting model to test.

1:01And this one, indeed, is pretty exciting. Ethan Molyc tweeted, I think we've been through enough release cycles for models at this point to say that the latest model from OpenAI or Anthropic or Google is generally going to be the best model in the world upon release, with some jagged edges until the next release by one of the big three. Now with that background, a different way to look at where we've been is that it's simply been OpenAI's turn. However, the expectations coming into GPT 5.4 were a little bit higher than they might have been for some of OpenAI's more recent releases. Ever since the release of GPT 5, all the big model providers got the memo that trying to promise too much in each update rather than just being very incremental was a pretty scary proposition. That's what got us the 5.1, 5.2, 5.3, now 5.4 kind of paradigm. But of course, it's not just OpenAI doing that, Google and Anthropic are both on that same plan as well. And yet, even with that, 5.4 has had a little bit more hype and anticipation around it than some of the previous iterative models that we've gotten more recently. This was theoretically supposed to be the big outcome of OpenAI's code red which was launched

2:02back in December. And what's more, the buzz for the last week or week and a half or so has been that this one was really meaningful. Enough so that it almost felt to me like some of the more recent leaks to publications like the information were almost trying to tamp down on expectations. To take one example, rumors had been flying that there was a 2 million token context window, whereas the information's reporting from the last couple of days suggested it was just 1 million. Seemed to me to be a little bit of expectation setting through leaks. In any case, on Thursday afternoon we actually got the model. And the initial buzz was strong. Ben Hylac wrote, I've been using GPT 5.4 for the past few weeks. In a sea of endless model drops and benchmark maxing, this model is the first and a long time to be worth your time to try. Honestly, didn't expect to open AI to pull this off. So let's talk first about how OpenAI frames things, look at some of the early reactions in the community, and then we'll walk through a more comprehensive case study with a project that I recently did to put the new capabilities through the ringer. Now one of the interesting things about this week is that this was not the first new OpenAI

3:04model we got. Just a couple of days ago we got GPT 5.3 instant, although OpenAI started promising almost immediately that 5.4 was coming. 5.3 instant was as we've talked about a speed and personality play. The announcement tweet called it more accurate, less cringe. This actually was part of the inspiration for our episode about what's going to actually matter in consumer AI, as this was so clearly aimed at that default sort of experience that the average chat GPT user is going to have because they're not optimizing the model selector for what they think they want. GPT 5.4, although still having a lot of offshoots, does feel like they're trying to bring together their models under a coherent banner. On the right, GPT 5.4 brings together the best of our recent advances in reasoning, coding and agent workflows into a single frontier model. It incorporates industry leading coding capabilities of GPT 5.3 codex, while improving how the model works across tools, software environments and professional tasks, involving spreadsheets, presentations and documents. The result is a model that gets complex real work done accurately, effectively and efficiently,

4:05delivering what you ask for with less back and forth. In other words, if 5.3 instant was for the personal use cases, GPT 5.4 was, as the subheader says, designed for professional work. As the information had it in that leak, we do indeed get a 1 million token context window, which among other things, improves 5.4's ability on tasks that require longer thinking. The quotes from early testers that they included were ravishing, even more than the ones that they usually include. Brendan Fudy, the CEO at Mirkoire Rights, GPT 5.4 is the best model we've ever tried. It's now top of the leaderboard on our Apex agents benchmark, which measures model performance for professional services work. It excels at creating long horizon deliverables such as slide decks, financial models and legal analysis, delivering top performance while running faster and at a lower cost than competitive frontier models. Indeed, these professional tasks are some of the big focus for open AI's announcement blog. They compare GPT 5.4's outputs to GPT 5.2's on spreadsheets, documents and presentations, all the significant upgrades. And if overall knowledge work is what open AI chose to focus on, there was also a big

5:08emphasis on computer use capabilities. We are, of course, living in an open-claw world, and so it makes sense that this is becoming increasingly important. Computer users actually listed above coding, perhaps because the jump from 5.3 codecs to 5.4 isn't all that big. Indeed, they basically said that 5.4 is the integration of 5.3 codecs with other improved aspects of the model, meaning it seems to me that their latest round of big coding innovations was embedded in 5.3 codecs itself. There's a sub-theme running throughout the announcement about efficiency, and the first part of the announcement posts they write that GPT 5.4 is our most token efficient reasoning model, using significantly fewer tokens to solve problems when compared to GPT 5.2, translating to reduce token usage and faster speeds. In the coding section, they also talk about fast mode in codecs, which they say delivers up to 1.5x faster token velocity. They write, it's the same model and same intelligence just faster. That means users can move through coding tasks, iteration, and debugging while staying in flow. The new efficiency gains also show up in tool search. They write, previously, when a model was given tools, all tool definitions were included

6:12in the prompt up front. For systems with many tools, this could add thousands or even tens of thousands of tokens to every request, increasing cost, slowing responses, and crowding the context with the information the model might never use. With tool search, GPT 5.4 instead receives a lightweight list of available tools, along with a tool search capability. When the model needs to use a tool, it can look up that tool's definition and append it to the conversation at that moment. This approach, they say, dramatically reduces the number of tokens required. They evaluated 250 tasks from scales, MCP atlas, and found that this new configuration had the same accuracy but reduced total token usage by 47%. That combined with improved accuracy on tool calling obviously makes it really appealing for agent use cases. There's a few other parts that the announcement, including improved web search and improved durability, but those are the big hits. And as people dug in, there were a few things that stood out. First of all, that efficiency is showing up in people's tests. Greg Kamrat, the president of ARC Prize, said that on ArcAGI 2, they were seeing a consistent 20% point lift versus 5.2 at the same price.

7:15Still the three benchmarks that people were most discussing were around coding, computer use, and GDPVAL. On coding people flagged that GPT 5.4 was only nominally better than 5.3 codecs on benchmarks like SWEBench Pro, but most people understood that that wasn't the core value proposition of this updated model. The computer use improvement got more of a discussion. This is maybe the most concerned I've ever seen people with computer use, and I think that's the difference between the pre and post open-claw world. Now that people have got all these Mac minis running around with their open-claw agents having pretty much unfettered access to them, how good the models are at using the computer becomes much more relevant in a day-to-day way, not just a theoretical way. Rahul Agarol writes, GPT 5.4 is here and it can use a computer better than a human. In AI, ship GPT 5.4 on March 5th. The headline isn't the reasoning improvements, it's that this is their first general-purpose model with native state-of-the-art computer use. It can operate websites and software autonomously, issue keyboard and mouse commands, write and execute code, and navigate full desktop environments.

8:16On OSworld verified it hits 75%, which is above human level performance at 72.4%, and a massive jump from GPT 5.2's 47.3%. That's not incremental, that's a step change. When agents can reliably navigate desktops, the bottleneck on automation shifts from can the model do it, to do you trusted enough to let it. That's the question nobody has a good answer to yet. Jimmy Cuff from PACE wrote an ex article about this specifically. He called it, we stress tested GPT 5.4 on the hardest UI on the internet, and writes, people don't realize how good AI computer use is actually gotten, until they see it tackle the hardest UIs in existence, legacy insurance portals. At PACE, we build AI agents for insurance workflows like submission and taken first notice of loss. Because we operate in this space, we use legacy enterprise software as our ultimate benchmark. If AI can reliably navigate a 20-year-old hyperdense insurance portal without hallucinating a click, it can navigate anything. For a long time, the technology just wasn't there, but after spending the last few months

9:17working closely with OpenAI to stress test their new GPT 5.4 model inside these environments, it became clear the paradigm has shifted. All things they point to as much better, the first is click accuracy. Historically, Jamie writes, the biggest failure point was simply clicking the right thing, enterprise software is incredibly dense, layouts are cluttered, buttons are tiny, and the systems were designed decades ago. Earlier models would frequently miss targets. GPT 5.4 is vastly better at grounding itself visually, clicking precisely where it needs to, even on a crowded screen. They also point out improvements in long trajectory reasoning, speed and time to iteration, as well as memory. Now there was also a lot of chatter around GDPVAL, basically around the measure of the model's performance on professional work tasks. GDPVAL remember tests against knowledge were expanding 44 occupations from the top nine industries that contribute to US GDP. The win rate versus industry professionals was 49.8% for GPT 5.2, 60% for 5.2 pro, and between 69.2% and 70.8% for the GPT 5.4 family.

10:19That's just wins when it comes to ties, that number rises to 82 to 83%. Ethan Mollock pointed out what this means in terms of time savings. He wrote, given the GDPVAL benchmark for GPT 5.4, the new model ties or beats humans is judged by other experts at professional tasks 82% of the time. If you give a seven hour task to AI, even with failure rates in the need to check results, you'd save four hours and 38 minutes on average. And in addition to the general performance increases across the GDPVAL set, it's very clear that open AI is aggressively going after certain industries even more. COO Brad Lightcap, for example, tweeted, the team worked extremely hard to make GPT 5.4 great for finance. It's much improved for financial modeling and analysis, integrates directly into Excel, and connects to Fectivia, Dilupa, S&P Global, and many more. It does feel like a codex moment is coming here. But what about the overall impressions outside of the individual benchmarks? Layton Space wrote, we've learned to take for granted that open AI is the smartest kid in the room, always reporting state-of-the-art eVALs, but this set of updates feels much more

11:21substantial and confident than any open AI launch in recent history. The every vibe check summed it up, open AI is back. Three months ago writes the team at every, open AI was losing the agent to coding race. Cloud code had captured developer's hearts, and Opus 4.5 was shipping at a level other models couldn't touch. Meanwhile, open AI's coding agent codex felt like it was built for an older era of coding with AI. It was precise but rigid, powerful but personalityless, and not good with tools are able to run for long periods of time autonomously. Open AI's latest model released GPT 5.4 along with their other recent releases GPT 5.3 codex, GPT 5.3 codex Spark, and the codex desktop app shifts the competitive balance back towards open AI on the coding front. The new model produces plans that are thorough and technically precise, and have a user-focused and human feel that has been missing from open AI's previous coding models. In our testing GPT 5.4 reviews code with more depth than GPT 5.3 codex, and has a noticeably more conversational voice. With a few tweaks, it became our preferred model to use in our open clause, especially

12:22given that it is half the price of Opus 4.6. Even Kieran Klassen, our diehard Cloud Code devotee, is now reaching for GPT 5.4 daily since we started testing it a week ago. As ever they say, there are trade-offs. GPT 5.4 has a tendency to expand the task well beyond what you asked for, and to call tasks done before they're finished. It sometimes completed tasks in obviously wrong ways, then lied about it. The bigger story here is Open AI's trajectory. From the codex desktop app to GPT 5.3 codex and to GPT 5.4, the company is iterating fast, and many members of the team now uses tools and models daily for coding, a significant change from a few months ago. A couple of things that they said they liked about it, it did proactive research without being asked, it had a more human voice than previous codexes, and it was roughly twice as fast as Opus. What they didn't like included scope creep on multi-step conversations, misreading completion reports in OpenClaw, and over-engineering. One of their team members called it too eager, adding things that aren't wrong, but aren't

13:22necessary. Ultimately, they summed it up with their title, Open AI is back. Other people who reported their own experience also had positive things to say. Dr. Daria Anute-Maz wrote, I asked GPT 5.4 pro to write a 10,000 word literary article on my beloved T-cells, as I read through it, I am both mesmerized and deeply emotional. It's just so beautifully written. When I was tweeting about what I found to be 5.4's over-verbosity, click Health, Simon Smith wrote, counterpoint, it's the best writing model from Open AI I've seen, and probably better than the best clawed models now at writing, and only needs a bit of nudging to write extraordinarily well. I've done a ton of writing tests with 5.4 now, and it's capable of writing every imaginable way with great empathy, creativity, wit, and decision, and it has personality again. Now IAI's Claire Voe wrote, what I love speaks more like a human than 5.2 and 5.3, a million token context window, tool use is Chef's Kiss, first model where I've experienced the true go-investigate-and-fix experience that feels robust. What needs work, basic frontend and UX taste, still loves a bullet point, latency, and stability.

14:25The next day she updated it and said she's, quote, starting to have loving feelings towards the model. Basically, it did some really difficult tech and data work really well without a lot of support. Matt Schumer, who you might remember from something big is happening, calls GPT 5.4, in short, the best model in the world by far. On coding, he writes, coding capabilities are ridiculous. It's essentially flawless. Inside codex, it's insanely reliable. Coding is essentially solved. There's not much more to say on this. It's just that good. Now, he did find weaknesses, including frontend taste, which is something that I'm going to get into in just a moment as well. Mark Tenon holds from perplexity also pointed out that while the model itself is great, the updates to the actual codex CLI experience are really good as well. He called them the real hero. So much less friction than the previous approval system, he says. Agenda AI is powering a $3 trillion productivity revolution, and leaders are hitting a real decision point. Do you build your own AI agents, buy off the shelf, or borrow by partnering to scale faster?

15:28KPMG's latest thought leadership paper, Agenda AI Untangled, navigating the build by or borrow decision, does a great job cutting through the noise or the practical framework to help you choose based on value, risk, and readiness. And how to scale agents with the right trust, governance, and orchestration foundation. Don't lock in the wrong model. You can download the paper right now at www.kPMG.us slash Navigate. Again, that's www.kPMG.us slash Navigate. There's a new standard that I think is going to matter a lot for the enterprise AI agent space. It's called AIUC1, and it builds itself as the world's first AI agent standard. It's designed to cover all the core enterprise risks, things like data and privacy, security, safety, reliability, accountability, and societal impact, all verified by a trusted third party. One of the reasons it's on my radar is that 11 Labs, who you've heard me talk about before and is just an absolute juggernaut right now, just became the first voice agent to be certified against AIUC1 and is launching a first of its kind, insurable AI agent. What that means in practice is real-time guardrails that block unsafe responses and protect

16:31against manipulation plus a full safety stack. This is the kind of thing that unlocks enterprise adoption. When a company building on 11 Labs can point to a third party certification and say our agents are secure, safe, and verified, that changes the conversation. Go to AIUC.com to learn about the world's first standard for AI agents. That's AIUC.com. Blitzie is driving over 5X engineering velocity for large-scale enterprises, a publicly traded insurance provider leveraged Blitzie to build a bespoke payments processing application, an estimated 13-month project, and with Blitzie, the application was completed in live in production in six weeks. A publicly traded vertical SaaS provider used Blitzie to extract services from a 500,000-line monolith without disrupting production 21 times faster than their pre-Blitzie estimates. These aren't experiments. This is how the world's most innovative enterprises are shipping software in 2026. You can hear directly about Blitzie from other Fortune 500 CTOs on the Modern CTO or CIO-classified podcasts. To learn more about how Blitzie can impact your SDLC, book a meeting with an AI solutions consultant at Blitzie.com, that's B-L-I-T-Z-Y.com.

17:37If you're an operator, your day is a non-stop stream of decisions, and most of them require you to look at the data. You don't need another dashboard. You need answers you can trust fast, but the bottleneck is always the same. The data isn't ready. It's scattered. It's messy. Definitions aren't clear. Depending on your data team or waiting on domain experts for clarification and confirmation, that's the bottleneck today's sponsor, PromptQL, is built to break. PromptQL is a trusted AI analyst for high-frequency decision-making. It connects across warehouses, databases, SAS, and internal APIs. No massive data prep or centralization required. It's built for multiplayer input. Teammates can jump into a thread, correct assumptions, and nuance, flag edge cases. PromptQL turns everyday conversations into a shared context. And if something is ambiguous, it doesn't guess. It escalates to the right expert, captures the correct logic, and gets it right next time. That's how it delivers trust inaccuracy. Over time, PromptQL specializes to your business, like that veteran employee who just knows things. From simple what is questions to complex what if scenarios, you can model impact and stress

18:39test decisions before you commit, all through a simple natural language prompt. PromptQL, the trusted AI analyst for teams with shared context and messy data. So as we round the corner, let's actually shift into the test that I did. I wanted to do something that was actually kind of difficult. And that was also real. I've been building a bunch of agents recently with Cloud Code, and the specific thing that I was interested in building with Codex and 5.4 was some type of experience to help people show off their agent building and agent orchestration skills. Right now, you know that we have a couple different self-directed experiences for people to learn how to use tools like OpenClaw. And one of the things that I've been recognizing around that is that this new skill set of agent building and agent orchestration is going to be more and more in demand from all sorts of different types of organizations. However, it is really this new types of skill set. It's not easy to describe necessarily. It's not easy to show necessarily. And the people who are going to be hiring or contracting for it aren't necessarily going

19:39to be really easily able to describe exactly even what they're looking for. They're just going to want people who can help them agentify things. And so I wanted to build some type of experience that could help that gap, that could help builders show off what they built in ways that were understandable and accessible to other people outside of them, and that could help people who were potentially interested in working with builders actually engage in a way where they weren't just totally blind in terms of finding the right people to help them with whatever agentification needs they had. So this is not an easy feat, and it also seemed to be a great way to test both 5.4 as a model inside Chatchee BT as well as the updated Codex. Now before we get into how it did, I will say that I also tested 5.3 instant as a partner in helping me get Codex set up on a new machine. Overall I think 5.3 instant is a big upgrade. It's less annoying than 5.2, which I could get into more, but honestly just 5.2 was really annoying. It wouldn't answer things and then it would be cloying in its responses. It's just a very unpleasant model to interact with. 5.3 is way better, it's much faster, and much easier to use for these not so intellectually

20:43heavyweight types of tasks. Two things that I have noticed with 5.3 instant however, one is that it puts too much into each response. When I said I need help setting up Codex and figuring out how to get the most out of it, it wrote this like 18 page tom with all sorts of steps basically giving me a full guide to Codex instead of working our way step by step. In fact one of the things that I would recommend if you are using it in this way is to ask it to go step by step. This over verbosity as you will see is one thing that I think is consistent with this new set of GPT models. I also noticed 5.3 instant doing this very weird clickbaity thing. Where when it has something to add at the end of a conversation, instead of just saying it, it gives you this weird almost clickbait type final sentence. For example, if you want I can also show you the 5 Codex workflows that technical founders are using to build products insanely fast right now. Later on it says, once that works I'll show you something important. The 32nd Codex workflow that makes it feel like you suddenly have a junior engineer working for you. It's like they studied at the school of clickbait linked in and X article headlines and

21:44brought that into chat GPT in a way that is just really really weird. Still over all those complaints aside, 5.3 instant is a big upgrade in a lot of ways especially for more simple types of needs. Now moving over into the planning stage, I used 5.4 thinking admittedly just on the standard setting and perhaps the results would have been very different if I had used more extended thinking or even pro. Even though this isn't the instant model, it is extremely fast compared to previous thinking iterations. In fact, it's so fast that it actually took some harnessing to slow it down and have it be actually useful. I described the problem that we were trying to solve in the same way that I described it to you and it immediately dove in trying to spec the thing out before we even discussed it. The problem was that this wasn't just over eager, it was that it fell into some default patterns based on its training data rather than really considering what I was trying to do differently. Specifically, in this case, it jumped right into assuming that if we were talking about matchmaking between a company and an agent orchestrator, we really needed to focus on technical skills.

22:44Even though, of course, the whole point is that we're dealing with a world where technical skills are being brought by the tools and so it's a different type of skill set that we're actually trying to understand. In addition to that, almost immediately I started to notice another issue, which is that 5.4 thinking was extremely oververbose. There's a point of repeating itself. Even when it's not required, like when it's agreeing with you about something that you already said, it uses a million lists, bullets, letter lists, numbered lists, all in the same response. And maybe this sounds like I'm complaining about it doing too much, which is maybe a better problem to have, but the way that it operates honestly puts a huge cognitive burden on the promsir. Now, like I said to be fair, I could have changed it from thinking standard to thinking extended to see if that slowed it down a little bit and improved things, but this was a set of issues that I was dealing with. And as I asked it to focus on our back-and-forth conversation, it was still trying to update its build spec at the same time. When it came to those biases that it was bringing into this conversation, that speed made it harder to undo. I didn't get it to slow down until I said, kind of annoyed, did I at any point ask you

23:48to immediately start building a spec? Finally, we got it to engage in the way that I wanted, and at that point I started running into a new set of challenges. One, it was still going incredibly long, and I found myself skipping over huge amounts of words that it was saying because there was just so much to read. The next thing that I noticed was that it was really eager to stay in planning mode, whereas way earlier in the conversation, Claude would have started building artifacts. GPT-5-4 just wanted to go deeper and deeper and deeper in planning in ways that I think were wildly over-optimizing, relative to just starting to build the thing. After one single line response where I told it what direction I wanted it to go, it agreed and then wrote hundreds of lines with multiple bulleted lists, just confirming what we had already talked about. Now, I will say that once we honed in on the core idea, it was pretty smart and was able to stay focused on the things that made what we were trying to build different, but I was still having real problems trying to get it to go into do the actual work mode. Even when I tried to shift it into visualization mode, it kept trying to plan the visualization rather than just visualize it.

24:49Again, I finally had to say, why aren't you just designing it? Claude would have showed me five versions by now. It responded, fair hit, I stayed too long an abstraction when I should have started showing. And then you would think that at that point it would start doing the showing, but instead it said, here are five concrete interface directions for builder rep, all built around blah, blah, blah, blah, blah, blah. I had to literally stop it and say, no, I'm not saying describe it, I'm saying go build the clickable prototype. Which it finally did, but then had its own problem, it was just awful visually. And this is something that lots of other folks pointed out. Ben Davis who works with Theo on his YouTube channel about AI said, it is hilariously bad at UI stuff. At Schumer said, front end taste is far behind Opus 46 and Gemini 3.1 Pro. Why is this so hard to fix? I'm not trying to rip on it too much here, but it is just honestly staggering. How bad and tasteless the UI design is in my little experience. I had to bring it back over to Claude and do the front end design there. Claude pulled no punches in critiquing what we had gotten from five four.

25:49The card backgrounds are muddy gradient blobs. The colors are dull and washed out. The typography has no hierarchy. The tags look cheap. The cards have no breathing room. The whole thing looks like a dark mode template from 2023. Brutal, but all very true. Still after a while, we use Claude to get the design under control, and we're off to the races inside the Codex CLI. And this is where things start to turn around, and where you start to see why many folks are so excited about this new model. There were some not perfect things about the experience with Codex. It had a couple weird foibles like it pushed me to use GPT 4.1, but I found similar weirdness like that with Claude code as well. I confirmed that it wasn't just five four in chat GPT, but also in Codex that was not good at design, but there were some parts of the experience that were great, including compared to Claude code. The thing that Mark Tenenholz was talking about with much less friction in the approval system was absolutely true. There are so much fewer confirmations with Codex right now than Claude code in ways that make the experience just massively, massively better. What's more, Codex did a much better job of letting you know what was going on as it

26:52was building. It actually has not just reasoning traces, but almost interstitial updates around what it's doing when it's doing a long running task. This means that when it's doing something that takes five or six or eight or ten minutes, it's not just a total black box. You actually have a sense of where it is in its process as it's happening. Now Claude code doesn't do none of that, but the way that Codex uses full sentences to describe where it is I think is really appealing. So the big thing was when we got to the point where we were actually ready to deploy this thing, there were zero errors. It just worked. It worked immediately, it worked right out of the box, in a way that basically nothing I've ever built with Claude code has. Where that leaves me is in a place that's pretty similar to what every found, where I'm not going to abandon one of these models for the other, but where I can absolutely see and probably will find five, four in Codex deeply integrated into my process going forward. I also think it's important to point out now that I've ripped on a bunch of things that I didn't like that this was my first set of tests, and I have not yet taken the time to go try to change, for example, the core instructions that make GPT 5.4 Overleaver

27:56Bose. It's entirely possible that a lot of my critiques are actually relatively superficial and fixable and just about dealing with the out of the box settings. Overall, based on the average of what people are experiencing, you would be doing yourself a disservice if you didn't go try GPT 5.4. I will certainly get more reps in with it, probably in some other contexts like OpenClaw as well, and I will report an update when I do. For now though, that is going to do it for today's AI Daily Brief, I appreciate you listening or watching, as always, have fun playing with 5.4, and until next time, peace!