OpenAI Retires GPT-4, Unveils New AI Models Amid Rapid Evolution

OpenAI is winding down its GPT-4 AI model in ChatGPT | TechCrunch

OpenAI will soon retire GPT-4, an AI model it launched over two years ago, from ChatGPT, the company announced in a changelog on Thursday. Effective April 30, GPT-4 will be "fully replaced" by GPT-4o, the current default model for ChatGPT, OpenAI said. However, GPT-4 will remain available in OpenAI's API. "In head‑to‑head evaluations, [GPT-4o] consistently surpasses GPT‑4 in writing, coding, STEM, and more," wrote OpenAI in the changelog. "Recent upgrades have further improved GPT‑4o's instruction following, problem solving, and conversational flow, making it a natural successor to GPT‑4." GPT-4 rolled out in March 2023 for ChatGPT, and also for Microsoft's Copilot chatbot on the web. Several versions of GPT-4 had multimodal capabilities, allowing them to understand both images and text -- a first for a widely deployed OpenAI model. OpenAI CEO Sam Altman has said that GPT-4, which is reportedly massive in size, cost more than $100 million to train. It was succeeded by GPT-4 Turbo in November 2023, a faster and cheaper model. GPT-4 is one of the models at the heart of copyright disputes between OpenAI and publishers including The New York Times. Publishers allege that OpenAI trained GPT-4 on their data without their knowledge or consent. OpenAI claims that fair use doctrine shields it from liability. GPT-4's coming retirement will likely follow the release of new models in ChatGPT. According to reverse engineer Tibor Blaho, OpenAI is readying a family of models called GPT-4.1 -- GPT-4.1 mini, GPT-4.1 nano, and GPT-4.1 -- as well as the o3 "reasoning" model the company announced in December and a new reasoning model called o4-mini.

OpenAI Switch-Up: We're Actually Getting ChatGPT o3 Soon

Katelyn is a writer with CNET covering social media, AI and online services. She graduated from the University of North Carolina at Chapel Hill with a degree in media and journalism. You can often find her with a novel and an iced coffee during her time off. OpenAI surprised folks this week when CEO Sam Altman announced on X/Twitter that the company is switching up its planned release schedule. The next ChatGPT model we can expect to see is o3 and o4-mini, coming soon in the next few weeks. But that means that GPT-5, the next major overhaul we expect for ChatGPT, is being pushed back to later this year. The news is a reversal of a February announcement that OpenAI was integrating o3 and other technology in the upcoming ChatGPT-5 model, meaning that o3 would no longer have its own release. In the meantime, you can use ChatGPT o3-mini. It's a smaller variant of o3, and we can take some clues from that model about what we can expect in o3. The o3-mini model is a reasoning model, meaning it "thinks" about prompts before responding and is better at solving complex problems. Free and paid ChatGPT users can try out o3-mini, which is also seen as a competitor to DeepSeek, the open-source AI that made a huge splash earlier this year. We don't know too much about o4-mini -- it's not the same as 4o-mini -- but it should be the next generation beyond o3. There are a number of other updates sure to keep people busy during the wait. We're also expecting the company to release its first open-weights model. OpenAI isn't very transparent about what goes on in the creation and training process of its models (despite its name), so the release of that model will be a big step as well. The open-weights, o3 and o4-mini models are the company's attempts to remain competitive amid tough competition, including Chinese AIs like DeepSeek. ChatGPT also got a native image generator, which was so popular that it temporarily overwhelmed the service's GPUs. Quickly, a trend of creating AI images mimicking the style of Studio Ghibli (the studio behind famous films such as Spirited Away and Ponyo) took off. Altman even changed his X/Twitter profile picture to a Studio Ghibli-esque version of himself. However, the founder of Studio Ghibli, Hayao Miyazaki, is notoriously against generative AI.

ChatGPT's GPT-4 model retires soon - some users can continue to access it

It seems like just yesterday that we were celebrating its release. When ChatGPT first gained popularity in 2022, GPT-3.5 was the default model. In March 2023, the more advanced GPT-4 replaced it. Since then, both models have been lapped by a much more capable GPT-4o, which will soon replace GPT-4. Also: ChatGPT will remember everything you tell it now - like a real personal assistant On Thursday, OpenAI updated its ChatGPT release notes to announce that GPT-4 was being retired from ChatGPT and fully replaced by GPT-4o starting April 30. In the changelog, OpenAI said that GPT-4o consistently outperformed its predecessor in writing, coding, and STEM in head-to-head evaluations. Retiring this model won't be much of a loss, as GPT-4o's release in May 2024 was a major upgrade. The "o" in GPT-4o stands for Omni and refers to the model's multimodal features. These features allow it to understand text, audio, image, and video inputs and output text, audio, and images, greatly expanding upon what GPT-4 could do. Since then, GPT-4o has continued to be upgraded to expand its assistance further. The changelog shows a March 27 upgrade that improved GPT-4o's capabilities across different areas, including smarter problem-solving in STEM and coding and enhanced instruction-following and formatting accuracy. OpenAI said these upgrades have made it a "natural successor" to GPT-4. Also: OpenAI is pushing for industry-specific AI benchmarks - why that matters Currently, subscribers can access GPT-4 under the model toggle, but they will be met with a caution symbol alerting them that the model is leaving on April 30. Developers and researchers will still be able to access GPT-4 in the API even past the retirement date, where they can also access other legacy models such as GPT-3.5, DALL-E 2, and more. This update won't impact free users as the default model for that model is GPT-4o mini, and the only other models they have limited access to include GPT-4o and o3-mini. The news comes at the same time as ChatGPT's latest feature update, which expanded the chatbot's memory feature to reference all of your past conversations and create a more personalized experience.

OpenAI gets ready to launch GPT-4.1

Tom Warren is a senior editor and author of Notepad, who has been covering all things Microsoft, PC, and tech for over 20 years. OpenAI is getting ready to unveil a number of new AI models, sources familiar with the company's plans tell The Verge. Among the new AI models will be a release of what I'm expecting will be branded GPT-4.1, which one source describes as a revamped version of OpenAI's GPT-4o multimodal model. GPT-4o was originally introduced last year as a flagship model that reasoned across audio, vision, and text in real time. I understand that OpenAI will launch GPT-4.1 alongside smaller GPT-4.1 mini and nano versions as soon as next week. OpenAI is also readying the full version of its o3 reasoning model and an o4 mini version that could debut even sooner. AI engineer Tibo Blaho discovered references to o4 mini, o4 mini high, and o3 in a new ChatGPT web version earlier today, suggesting these additions are imminent. I understand o3 and o4 mini are both set to debut next week, unless OpenAI moves the launch plans around. OpenAI CEO Sam Altman teased on X that OpenAI would be launching an exciting feature today, but it's not clear if this is related to the o3 and o4 mini references in ChatGPT or not. Sources caution that OpenAI has delayed the introduction of some new models recently due to capacity issues, so it's possible for the new GPT-4.1 model introduction to slip beyond a planned debut next week. I asked OpenAI to comment on this story, but the company didn't respond in time for publication. Altman revealed on X earlier this month that customers "should expect new releases from OpenAI to be delayed, stuff to break, and for service to sometimes be slow as we deal with capacity challenges." OpenAI's more advanced image generation capabilities forced the company to temporarily rate limit requests last month, and Altman claimed "our GPUs are melting" due to the popularity of the built-in image generator for users of ChatGPT's free tier.

Two Years After GPT-4 Broke the Internet, OpenAI Is Quietly Killing It

OpenAI will retire the once-groundbreaking model by month's end. GPT-4o is the new flagship and GPT-5 is waiting in the wings. OpenAI is retiring GPT-4, one of its most well-known AI models. Effective April 30, GPT-4 will be removed from the drop-down menu known as the "model picker" for ChatGPT Plus users. It will remain an option in the API. Those who are using ChatGPT for free cannot select the model they use, so their experience won't change. "GPT‑4 marked a pivotal moment in ChatGPT's evolution," OpenAI says. "We're grateful for the breakthroughs it enabled and for the feedback that helped shape its successor." When GPT-4 debuted in March 2023, it was a noticeable improvement over ChatGPT's initial model, GPT-3.5. OpenAI spent over $100 million training GPT-4, TechCrunch reports. Some of that training included copyrighted materials, The New York Times argues in an ongoing lawsuit. Two years later, the GPT-4o model debuted as the new flagship. Today, it "consistently surpasses GPT‑4 in writing, coding, STEM, and more," and delivers "even greater capability, consistency, and creativity," OpenAI says. Other available models include GPT-4 Turbo, GPT-o3, GPT-o1, and GPT-4.5, which launched this year. There are also rumors of an upcoming GPT-4.1 model. Having that many models can be confusing, OpenAI CEO Sam Altman admitted in February. He hopes to eventually do away with the model picker, which puts the onus on the user to select the appropriate model for the task, and move toward a "unified intelligence" solution, where the system picks. All the models it offers are experiments and steps toward its next big release, GPT-5, which has proven difficult to achieve. OpenAI is battling GPU shortages and those graphics cards aren't cheap. Last month, it closed a massive $40 billion funding round, which it says will "push the frontiers of AI research even further, scale our compute infrastructure, and deliver increasingly powerful tools." That will, presumably, include GPT-5. When GPT-5 arrives, Altman says it will be available to free users on a "standard intelligence" setting. Plus subscribers ($20/month) will be able to run it at "a higher level of intelligence," and Pro subscribers ($200/month) at an "even higher level of intelligence" that includes Voice mode, Canvas, Search, Deep Research, and other OpenAI products.

ChatGPT's o4-mini, o4-mini-high and o3 spotted ahead of release

OpenAI is preparing to launch as many as three new AI models, possibly called "o4-mini", "o4-mini-high" and "o3". Right now, ChatGPT has as many as five models, including GPT 4o (the non-reasoning model), GPT 4.5 (another non-reasoning model but with greater creativity), and three reasoning models: o1, o3-mini, and o3-mini-high. o1's successor is o3, but a full-fledged model isn't available yet. We only have access to the o3-mini and o3-mini-high, which are small reasoning models in the o-series, offering improved performance and faster responses. According to references found on ChatGPT's web app, OpenAI is preparing to launch as many as three new models: o3 (a full-fledged reasoning model in the o3-mini, o3-mini-high lineup), and o4-mini, o4-mini-high, similar to the existing models but with greater reasoning capabilities. In a post on X last week, OpenAI CEO Sam Altman previously confirmed that OpenAI plans to launch new o3 and o4 models before the upcoming GPT-5, which is said to be a unified reasoning model. "We are going to release o3 and o4-mini after all, probably in a couple of weeks, and then do GPT-5 in a few months," Sam Altman wrote in a post on X. "There are a bunch of reasons for this, but the most exciting one is that we are going to be able to make GPT-5 much better than we originally thought. We also found it harder than we thought it was going to be to smoothly integrate everything, and we want to make sure we have enough capacity to support what we expect to be unprecedented demand." We don't know when OpenAI plans to ship these three new models, but preparations are clearly underway based on the references.

ChatGPT 4.1 could launch as soon as next week - here's why that's a big deal

An exclusive from The Verge reveals sources close to OpenAI have confirmed that the next major ChatGPT release is just around the corner. The report claims "one source describes [the new model] as a revamped version of OpenAI's GPT-4o multimodal model." The source claims OpenAI will launch "GPT-4.1 alongside smaller GPT-4.1 mini and nano versions as soon as next week." This release would mark a major step forward for ChatGPT. Over the last few months, Google Gemini and DeepSeek have continuously improved, closing the gap with OpenAI's flagship model. You'd expect these new models to propel OpenAI back to the top spot. AI engineer Tibor Blaho found code in the new ChatGPT web app version mentioning "o4-mini", "o4-mini-high" and "o3" which would imply the new GPT models are likely to arrive very soon. Just yesterday, OpenAI CEO, Sam Altman, announced major memory upgrades coming to ChatGPT which means the AI chatbot can remember every conversation it has with a user. This upgrade, which is likely to be part of any new GPT means ChatGPT can now take that next step to becoming your personal assistant. Whether that's referencing previous conversations or the AI having full awareness of your past interactions, a new memory upgrade is a major improvement coming to ChatGPT. The name of the incoming update is not confirmed, although The Verge believes it will be called GPT 4.1. If OpenAI goes down the 4.1 route, it will continue to make the AI giant's model naming scheme incredibly convoluted. At the end of last year, OpenAI mentioned that the company hoped to create an all-encompassing model in the future, which would not require the user to select between models. If ChatGPT 4.1 releases next week, we'd really be hoping to see that all-encompassing model sooner rather than later, as trying to explain the differences between 4.1, 4.5, and o3 would start to get very complicated indeed. Naming aside, a new OpenAI AI model is a huge deal, and if these sources are correct and we see a new launch in the next week or so, it could be another major step towards the AI future these tech CEOs are pursuing.

OpenAI is retiring GPT-4 from ChatGPT

OpenAI's GPT-4 model will soon be "fully replaced" by its natively multimodal GPT-4o model. According to ChatGPT's release notes (via TechCrunch), "GPT-4 will be retired from ChatGPT" on April 30. The model, which was released over two years ago, will still be available in the API, but recent updates to GPT-4o have rendered GPT-4 somewhat obsolete. "Recent upgrades have further improved GPT‑4o's instruction following, problem-solving, and conversational flow, making it a natural successor to GPT‑4," the note read. For those who have been following OpenAI and the AI industry, it puts the breakneck speed of the industry into sharp relief, while simultaneously pointing out that GPT-5 has yet to emerge. GPT-4, released in March 2023, was a notable step up from GPT-3.5, the previous model, which ushered in the ChatGPT explosive introduction to the world. Since then, OpenAI has released the multimodal GPT-4o and iterations of its "o-series" models, which have "chain-of-thought" reasoning capabilities. This announcement isn't entirely unexpected. In March, OpenAI CEO Sam Altman shared on X that the company was "simplifying" its offerings and would make GPT-4.5 its last "non-chain-of-thought model." After that, the company would focus on combining its o-series and GPT models into a product that's trained to adapt to users' needs. Making GPT-4o the default model is a step towards that effort to streamline its products. This Tweet is currently unavailable. It might be loading or has been removed. Currently, OpenAI plans to release o3 and o4-mini as standalone models and has pushed back the GPT-5 release, which Altman says will ship "in a few months." That all could change though, since GPT-5 has been delayed several times now. This Tweet is currently unavailable. It might be loading or has been removed. However, according to the release note, GPT-4 has officially served its purpose as a "pivotal moment in ChatGPT's evolution." Its services are no longer required. RIP.

ChatGPT-5 is on hold as OpenAI changes plans and releases new o3 and o4-mini models

The news broke today in a tweet by OpenAI CEO Sam Altman, in which he revealed why the plans were changing: "There are a bunch of reasons for this", wrote Altman, "but the most exciting one is that we are going to be able to make GPT-5 much better than we originally thought. "We also found it harder than we thought it was going to be to smoothly integrate everything, and we want to make sure we have enough capacity to support what we expect to be unprecedented demand." The mention of 'capacity to support unprecedented demand' is clearly a reference to the recent outages that ChatGPT has been experiencing as millions of new users signed up to try out the new image generation abilities of ChatGPT-4o. ChatGPT-5 is the next big evolution of the popular ChatGPT LLM and will be a major development in the future of AI. Its simpler name was also supposed to represent an alignment shift in OpenAI's somewhat confusing product-naming conventions that will now soon feature both an o4 and an 4o model in the line-up simultaneously. Rather than the user having to decide if they wanted to use a smaller, lighter model, such as 4o-mini or a deeper reasoning model, like o4, for their tasks, ChatGPT-5 will decide for you which type of model to use, based on your query. So far, OpenAI has confirmed that even users on the free tier will have some access to ChatGPT-5 when it comes out, but users on the Pro and Plus tiers will get more. The only word on a release date we've been given before was "soon". Now it looks like we'll have to wait a little bit longer for that integration of everything into one model, with Altman stating that ChatGPT-5 would now appear "in a few months". Commenting on the new o3 model, Altman also stated that, "We were able to really improve on what we previewed for o3 in many ways; I think people will be happy..." Replying to a user on X who asked if there would also be an o3 Pro model, Altman gave a one-word reply - "coming!" - which would seem to confirm that a pro version of o3 is also in the works. As to when we will see the o3 and o4-mini models, Altman stated, "in a couple of weeks, and then do GPT-5 in a few months".

OpenAI adjusts AI roadmap for better GPT-5

OpenAI is reconfiguring its rollout plan for upcoming AI models. The company's CEO, Sam Altman shared on social media on Friday that it will delay the launch of its GPT-5 large language model (LLM) in favor of some lighter reasoning models to release first. The brand will now launch new o3 and o4-mini reasoning models in the coming weeks as an alternative to the GPT-5 launch fans were expecting. In this time, OpenAI will be smoothing out some issues in developing the LLM before a final rollout. The company hasn't detailed a specific timeline, just indicating that GPT-5 should be available in the coming months. Recommended Videos "There are a bunch of reasons for this, but the most exciting one is that we are going to be able to make GPT-5 much better than we originally thought," Altman said in an X post. "We also found it harder than we thought it was going to be to smoothly integrate everything, and we want to make sure we have enough capacity to support what we expect to be unprecedented demand," he added. Prior reports suggested that GPT-5 might have been prepared for release in the May timeframe; however, several unforeseen developments have popped up since then. TechRadar noted that OpenAI is likely having to tackle the tons of new users its ChatGPT services has recently acquired. Its user base recently jumped from 400 million to 500 million in a matter of hours, after a design trend prompted by its latest GPT-4o image generation update went viral. While the GPT-5 update has been long anticipated, the incremental updates are expected to help set up the introduction of the major rollout. The publication noted that once the o3 and o4-mini models are available, OpenAI will have products called o4 and 4o within the ChatGPT ecosystem. By the time the GPT-5 rolls out, it will have the capability to select the best model for your task.

OpenAI Plans O3 and O4-Mini Release Before GPT-5, Altman Says - Decrypt

OpenAI CEO Sam Altman revealed a significant shift in the company's release plans on Friday, announcing that two intermediate models will arrive before its highly anticipated GPT-5. "Change of plans: We are going to release o3 and o4-mini after all, probably in a couple of weeks, and then do GPT-5 in a few months," Altman wrote on X Friday. The surprise announcement comes as OpenAI grapples with technical complexities in its flagship model development. Altman admitted the company "found it harder than we thought it was going to be to smoothly integrate everything" into GPT-5, suggesting the staggered release will help ensure sufficient capacity "to support what we expect to be unprecedented demand." The move places OpenAI in an increasingly crowded field of AI heavyweights rolling out advanced models. Google recently launched Gemini 2.5 Pro, which boasts 1 million tokens of context and has been widely regarded as the best reasoning and coding model available -- and is free to use. Meanwhile, DeepSeek R2, Grok-3, and Claude 3.7 Sonnet with extended thinking capabilities are all slated for imminent release -- each undercutting OpenAI's reasoning model on price. Altman teased a silver lining: releasing the intermediate models will give OpenAI more time to supercharge GPT-5. "The most exciting [reason] is that we are going to be able to make GPT-5 much better than we originally thought," he wrote. GPT-5 is expected to be fully multimodal, merging all of OpenAI's specialized models into a single system. That would eliminate the current need for ChatGPT to switch between reasoning models, standard language models, and image generation models based on the prompt. Instead, all these functions would be handled by a unified model. Technical specifications for o3 and o4-Mini remain under wraps, but they're expected to bridge the capabilities gap between GPT-4 and the forthcoming GPT-5, which industry watchers believe will feature substantial improvements in reasoning, planning, and memory functions. OpenAI's latest release, the reasoning-focused o1 Pro, came with eyebrow-raising pricing: $150 per million tokens (~750,000 words) for input, and $600 per million tokens generated. That's double the input cost of GPT-4.5 and 10 times the price of regular o1. And for reference, DeepSeek R1 costs less than $1 per million tokens. The revised roadmap arrives just days after OpenAI closed a historic $40 billion funding round -- the largest single fundraising event by any private tech company.

OpenAI Shifts Course, Says GPT-5 Coming in 'a Few Months'

Currently, OpenAI has two primary lines of models: GPT models and O models. The GPTs are traditional large language models, like GPT-4o or GPT-4.5, while the Os are reasoning models, meaning they can think through (to the extent that an AI is actually thinking) how to solve a problem in multiple steps. In February, Altman posted on X that the two lines would be combined into a single line of powerful AI models, starting with GPT-5. At the time, Altman wrote that the company's newest O model, o3, would not be released as a standalone product and would instead be integrated into GPT-5. But now, Altman has reversed course. He said that OpenAI will release o3 in "a couple of weeks," and will also release o4-mini, a new more affordable reasoning model, in a similar timeframe.

OpenAI GPT-5 delay confirmed, but new models are coming soon - Phandroid

The OpenAI GPT-5 delay is official. Sam Altman just broke the news, and if you've been waiting since GPT-4o, yeah, it stings a little. GPT-5 won't be landing anytime soon, but there's still plenty to look forward to. Altman explained the delay was due to "a bunch" of reasons. OpenAI's been having trouble integrating the new model, and they're bracing for what he calls "unprecedented demand." Right now, OpenAI is already feeling the heat. Ever since they upgraded ChatGPT with image generation, the servers have been slammed. People love visual tools, and the latest update made them faster, sharper, and way more reliable. It's no surprise that it's causing a traffic spike. Then there's ChatGPT's new Deep Research tool. This is a game-changer for anyone using AI to dig for info. Deep Research lets you tackle complex topics with better sources and longer answers. Think of it as ChatGPT with a research assistant baked in. OpenAI is clearly building momentum, even if GPT-5 isn't ready yet. And it's not like we're left empty-handed while we wait. OpenAI is rolling out two new models: o3 and o4-mini. These are focused on reasoning and should bring a noticeable step up in performance. Altman teased that they'll land in just a few weeks, giving us something to chew on before the big one drops. So yes, the OpenAI GPT-5 delay is a bit of a bummer. But with smarter tools and fresh models on the way, the hype isn't slowing down anytime soon.

OpenAI to replace GPT-4 with GPT-4o on ChatGPT from April 30

GPT-4 will continue to be available for developers via OpenAI's API. The Microsoft-backed company said it has introduced a memory feature in ChatGPT, which allows the chatbot to reference previous user conversations. This is aimed at making responses more personalised and context-aware.Microsoft-backed OpenAI will remove its GPT-4 model from ChatGPT on April 30, the company said in a changelog update posted on Thursday. The artificial intelligence (AI) firm will fully transition to GPT-4o, which is now the default model on ChatGPT. "Effective April 30, GPT-4 will be fully replaced by GPT-4o," the company said. GPT-4 will continue to be available for developers via OpenAI's application programming interface (API). OpenAI said GPT-4o outperforms GPT-4 across a range of tasks. "In head‑to‑head evaluations, (GPT-4o) consistently surpasses GPT‑4 in writing, coding, STEM, and more," it said. "Recent upgrades have further improved GPT‑4o's instruction following, problem solving, and conversational flow, making it a natural successor to GPT‑4." OpenAI had introduced GPT-4 in March 2023. It was the first widely deployed OpenAI model with multimodal capabilities -- the ability to process both text and images -- and it powered both ChatGPT and Microsoft's AI assistant Copilot. OpenAI chief executive officer Sam Altman has previously said GPT-4 cost more than $100 million to train. GPT-4 is also part of several copyright lawsuits against OpenAI, including one filed by The New York Times. The publishers allege that OpenAI used their content to train GPT-4 without permission. OpenAI has denied wrongdoing, saying its use of publicly available data is protected under the fair use doctrine. Meanwhile, the company has introduced a memory feature in ChatGPT, which allows the chatbot to reference previous user conversations. The move is aimed at making responses more personalised and context-aware. "We have greatly improved memory in ChatGPT -- it can now reference all your past conversations!" Altman posted on X.

OpenAI to replace GPT-4 with GPT-4o from April 30

GPT-4 will continue to be available for developers via OpenAI's API. The Microsoft-backed company said it has introduced a memory feature in ChatGPT, which allows the chatbot to reference previous user conversations. This is aimed at making responses more personalised and context-aware.Microsoft-backed OpenAI will remove its GPT-4 model from ChatGPT on April 30, the company said in a changelog update posted on Thursday. The artificial intelligence (AI) firm will fully transition to GPT-4o, which is now the default model on ChatGPT. "Effective April 30, GPT-4 will be fully replaced by GPT-4o," the company said. GPT-4 will continue to be available for developers via OpenAI's application programming interface (API). OpenAI said GPT-4o outperforms GPT-4 across a range of tasks. "In head‑to‑head evaluations, (GPT-4o) consistently surpasses GPT‑4 in writing, coding, STEM, and more," it said. "Recent upgrades have further improved GPT‑4o's instruction following, problem solving, and conversational flow, making it a natural successor to GPT‑4." OpenAI had introduced GPT-4 in March 2023. It was the first widely deployed OpenAI model with multimodal capabilities -- the ability to process both text and images -- and it powered both ChatGPT and Microsoft's AI assistant Copilot. OpenAI chief executive officer Sam Altman has previously said GPT-4 cost more than $100 million to train. GPT-4 is also part of several copyright lawsuits against OpenAI, including one filed by The New York Times. The publishers allege that OpenAI used their content to train GPT-4 without permission. OpenAI has denied wrongdoing, saying its use of publicly available data is protected under the fair use doctrine. Meanwhile, the company has introduced a memory feature in ChatGPT, which allows the chatbot to reference previous user conversations. The move is aimed at making responses more personalised and context-aware. "We have greatly improved memory in ChatGPT -- it can now reference all your past conversations!" Altman posted on X.

OpenAI to release o3, o4-mini ahead of GPT-5 launch later this year

The shift in the company's AI model release schedule is an attempt to maintain its competitive edge as it faces growing competition in the open-source space from Meta and Chinese rival DeepSeek.OpenAI CEO Sam Altman announced a shift in the company's artificial intelligence (AI) model release schedule, revealing that intermediate models known as 'o3' and 'o4-mini' will be launched in the coming weeks, ahead of the much-anticipated GPT-5 rollout scheduled for later this year. In a tweet on Thursday, Altman cited several reasons for the change, including unexpected technical challenges and a desire to better prepare for what he described as "unprecedented demand" for GPT-5. He also hinted that the new model would surpass initial expectations, stating, "The most exciting reason is that we are going to be able to make the GPT-5 much better than we originally thought." The comment came after OpenAI CEO Sam Altman on Wednesday hinted at enhancements to ChatGPT's recently launched image-generation tool, which has seen a significant surge in recent days. ChatGPT has crossed 130 million users, generating over 700 million images since March 27. The move is an attempt to maintain its competitive edge as it faces growing competition in the open-source space from Meta and Chinese rival DeepSeek. The Chinese low-cost model, DeepSeek, was developed in two months, with an investment of less than $6 million. This starkly contrasts the $100 million OpenAI reportedly spent on training its GPT-4 model. DeepSeek's R1 model offers developers, researchers, and organisations seeking AI solutions $0.55 per million input tokens and $2.19 per million output tokens. For comparison, OpenAI charges $15 per million input tokens and $60 per output token, ET reported in January.

Sam Altman Backtracks On o3 Cancellation, Plans To Release Model In Anticipation Of GPT-5: 'People Will Be Happy'

Feel unsure about the market's next move? Copy trade alerts from Matt Maley -- a Wall Street veteran who consistently finds profits in volatile markets. Claim your 7-day free trial now. OpenAI, led by CEO Sam Altman, has spent the past two years in fierce competition with rivals Google Gemini, Anthropic, DeepSeek and Elon Musk's xAI. Altman announced Friday that the San Francisco-based company's model release schedule will undergo an important change. What Happened: OpenAI had intended to shelve its o3 and o4 models earlier this year, planning to integrate them into the company's upcoming GPT-5. At the time, Altman said he wanted to simplify OpenAI's model offerings. On Friday, Altman announced that the company was reversing course, anticipating the release of o3 and o4 in a "couple of weeks" before rolling out the long-awaited GPT-5 in 2025. "we were able to really improve on what we previewed for o3 in many ways; i think people will be happy...," Altman said in a reply. The entrepreneur believes GPT-5 will be "much better" than originally thought, but the company will need additional capacity to support user demand. Why it Matters: OpenAI has repeatedly pushed back the release of GPT-5. While the company's flagship ChatGPT is the most popular AI model in the world, competitors have rapidly improved their offerings. Anthropic, a Jeff Bezos-backed AI startup, one-upped OpenAI in some areas in its update of Claude in February. DeepSeek, a Chinese company, released its R1 model in January. Also Read: JPMorgan Raises Recession Risk To 60% As 'Largest US Tax Hike' In 60 Years Hits Global Economy Photo: Shutterstock Market News and Data brought to you by Benzinga APIs

OpenAI Delays GPT-5 But Sam Altman Says o3 and o4-mini Models are Coming Soon

In addition, OpenAI will also release the o3-pro model which takes more time to think to solve harder problems. In February this year, OpenAI CEO Sam Altman shared the roadmap for GPT-5 release. He revealed that the powerful o3 reasoning model will not be available as a standalone model, instead, it will be integrated into GPT-5. Now, Altman says that o3 and the next-gen o4-mini reasoning models are coming "in a couple of weeks", and GPT-5 will be released "in a few months." On Friday, Altman tweeted on X, "change of plans: we are going to release o3 and o4-mini after all, probably in a couple of weeks, and then do GPT-5 in a few months." Additionally, OpenAI will be releasing an o3-pro model, just like o1-pro, that uses more time to solve harder problems. Altman has explained the reasons behind the GPT-5 delay. He says that integrating a lot of tools into GPT-5 including the o3 model is "harder than we thought." In addition, GPT-5 is going to be improved in the meantime and OpenAI is building the infrastructure and capacity to serve the GPT-5 model when it's released in a few months. In earlier statements, Altman has said that GPT-5 will be available to free ChatGPT users as well. Given that Studio Ghibli-style images blew up on ChatGPT, it's good that OpenAI is focusing on building capacity. Apart from that, the early release of o3 may have to do with Google's flagship Gemini 2.5 Pro model which has positioned itself as the leading AI model in the industry, dethroning all competing models. OpenAI may want to launch its impressive o3 model to take the lead against Google and Anthropic.