ChatGPT’s newest GPT-4 upgrade makes it smarter and more conversational

Developers should brace themselves—not just for cheaper AI, but for the domino effect this pricing revolution might trigger as Anthropic, Google, and xAI scramble to keep pace. For teams previously limited by cost, complexity, or both, GPT-4.1 might just be the catalyst for a new wave of AI-powered innovation. Moreover, Gemini lacks an automatic billing shutdown, which Prompt Shield says exposes developers to Denial-of-Wallet attacks—malicious requests designed to deliberately inflate your cloud bill, which Gemini’s current safeguards don’t fully mitigate. GPT-4.1’s predictable, no-surprise pricing seems to be a strategic counter to Gemini’s complexity and hidden risks. Even if GPT-5 doesn’t reach AGI, we expect the upgrade to deliver major upgrades that exceed the capabilities of GPT-4. I have been told that gpt5 is scheduled to complete training this december and that openai expects it to achieve agi.

Tech and VC heavyweights join the Disrupt 2025 agenda

In a follow-up response to VentureBeat, OpenAI communications confirmed that the June email was simply a scheduled reminder and that there are currently no plans to remove GPT-4.5 from ChatGPT subscriptions, where the model remains available. Despite the strong reaction, OpenAI had in fact already announced the plan to deprecate GPT-4.5 Preview back in April 2025 during the launch of GPT-4.1. Some described GPT-4.5 as a daily tool in their workflow, praising its tone and reliability.

Like previous GPT models from OpenAI, GPT-4 was trained using publicly available data, including from public web pages, as well as data that OpenAI licensed. At that time, the company stated that developers would have three months to transition away from 4.5. OpenAI framed the model as an experimental offering that provided insights for future development, and said it would carry forward learnings from GPT-4.5 into future iterations — particularly in areas like creativity and writing nuance. The GPT-4.1 models should help software engineers who are using ChatGPT to write or debug code, OpenAI spokesperson Shaokyi Amdo told TechCrunch.

We also tried generating images using the current image generator in Meta AI and the results were disappointing to say the least, especially when compared to ChatGPT’s new feature. But the addition of GPT-4.1 to ChatGPT appears to contradict that simplification goal. 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. 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.”

Windsurf bets big on GPT-4.1’s developer appeal

A lot has changed since then, with Microsoft investing a staggering $10 billion in ChatGPT’s creator OpenAI and competitors like Google’s Gemini threatening to take the top spot. Since then, we’ve seen GPT-4 evolve significantly, spanning several related models. It’s only a matter of time before GPT-5 arrives, but what exactly can we expect when it does? To answer that question, let’s dive right in and take a look at everything we know or have heard rumored about GPT-5. OpenAI’s GPT-4.1 isn’t just shaking up the pricing game, it’s potentially setting new standards for the AI development community.

The image-understanding capability isn’t available to all OpenAI customers just yet. But OpenAI, for its part, is forging full steam ahead — evidently confident in the enhancements it’s made. OpenAI does note, though, that it made improvements in particular areas; GPT-4 is less likely to refuse requests on how to synthesize dangerous chemicals, for one. Pricing is $0.03 per 1,000 “prompt” tokens (about 750 words) and $0.06 per 1,000 “completion” tokens (again, about 750 words). Tokens represent raw text; for example, the word “fantastic” would be split into the tokens “fan,” “tas” and “tic.” Prompt tokens are the parts of words fed into GPT-4 while completion tokens are the content generated by GPT-4. Judging by the graphs provided, the biggest jumps in capabilities are in mathematics and GPQA, or Graduate-Level Google-Proof Q&A – a benchmark based on multiple-choice questions in various scientific fields.

• There’s no public roadmap for GPT-5 yet, but OpenAI might have an intermediate version ready in September or October, GPT-4.5.
• Some notable personalities, including Elon Musk and Steve Wozniak, have warned about the dangers of AI and called for a unilateral pause on training models “more advanced than GPT-4”.
• March 31, 2023 – Italy banned ChatGPT for collecting personal data and lacking age verification during registration for a system that can produce harmful content.
• This isn’t mere generosity; it’s a strategic gamble that once developers experience GPT-4.1’s capabilities and cost savings firsthand, reverting to pricier or less capable models will be a tough sell.
• The tech giant has been modifying its Copilot experience, adding extra features over time while sticking its chatbot UI in every 365 app, Github, Dropbox, and into Windows 11 itself.

John-Anthony Disotto is TechRadar’s Senior Writer, AI, bringing you the latest news on, and comprehensive coverage of, tech’s biggest buzzword. An expert on all things Apple, he was previously iMore’s How To Editor, and has a monthly column in MacFormat. John-Anthony has used the Apple ecosystem for over a decade, and is an award-winning journalist with years of experience in editorial. 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.

That generous 75% caching discount, effectively incentivizing developers to optimize prompt reuse—particularly beneficial for iterative coding and conversational agents. Finally, once GPT-5 rolls out, we’d expect GPT-4 to power the free version of ChatGPT. AGI is best explained as chatbots like ChatGPT becoming indistinguishable from humans. AGI would allow these chatbots to understand any concept and task as a human would. Since GPT-4 is such a massive upgrade for ChatGPT, you wouldn’t necessarily expect OpenAI to be able to significantly exceed the capabilities of GPT-4 so soon with the upcoming GPT-5 upgrade.

The next generation of OpenAI’s GPT models could launch as soon as next week, with a new report hinting at the arrival of ChatGPT 4.1 to replace 4o. Even with system messages and the other upgrades, however, OpenAI acknowledges that GPT-4 is far from perfect. It still “hallucinates” facts and makes reasoning errors, sometimes with great confidence. In one example cited by OpenAI, GPT-4 described Elvis Presley as the “son of an actor” — an obvious misstep. A more meaningful improvement in GPT-4, potentially, is the aforementioned steerability tooling.

GPT-4 was a major update

Last year, Microsoft debuted the Copilot+ line of PCs with their set standard for on-device AI processing. The tech giant has been modifying its Copilot experience, adding extra features over time while sticking its chatbot UI in every 365 app, Github, Dropbox, and into Windows 11 itself. In theory, Meta states that its Llama 4 Maverick model has an upper hand over OpenAI’s GPT-4o model in coding, reasoning, multilingual, long context, and image benchmarks. In the LMarena leaderboard – at the time of writing, Llama 4 Maverick leads both GPT-4o and GPT-4.5 Preview, and is in second place behind the Gemini 2.5 Pro (Experimental) model. However, it is important to note that these benchmarks are often fluid and results can change as new updates are rolled out.

Benchmarks:

Additionally, ChatGPT recently got a Deep Research mode that allows the chatbot to perform a rigorous search of the web to find what OpenAI claims are research assistant level answers. While the end results may not always match the company’s claims, it’s still an important tool to have in order to conduct a comprehensive search of the web, as evidenced by the adoption by Perplexity AI, Grok, and Gemini. Reddit users and OpenAI forum members alike commonly voice confusion about the available options. “I do not understand the reason behind having multiple models available for use,” wrote one Reddit user in March. “Why would anyone use anything but the best one?” Another Redditor said they were “a bit lost” with the many ChatGPT models available after switching back from using Anthropic Claude.

Based out of Stockport, England, on TechRadar you’ll find him covering news, features and reviews, particularly for phones, tablets and wearables. Recently, OpenAI added native image generation capabilities to GPT-4o, allowing the model to generate more accurate and nuanced images, leading to a craze for viral Studio Ghibli-style images, among other trends. May 13, 2024 – A big day for OpenAI, when the company introduced the GPT-4o model, offering enhanced intelligence and additional features for free users. April 25, 2023 – OpenAI added new ChatGPT data controls that allow users to choose which conversations OpenAI includes in training data for future GPT models. February 1, 2023 – OpenAI announced ChatGPT Plus, a premium subscription option for ChatGPT users offering less downtime and access to new features.