ChatGPT: Everything you need to know about the AI-powered chatbot
- By Matthew Terry
- March 14, 2025
- Comments
- 30 minutes read
- 367 Views
- 8 months ago
ChatGPT, OpenAI’s text-generating AI chatbot, has taken the world by storm since its launch in November 2022. What started as a tool to supercharge productivity through writing essays and code with short text prompts has evolved into a behemoth with 300 million weekly active users.
2024 was a big year for OpenAI, from its partnership with Apple for its generative AI offering, Apple Intelligence, the release of GPT-4o with voice capabilities, and the highly-anticipated launch of its text-to-video model Sora.
OpenAI also faced its share of internal drama, including the notable exits of high-level execs like co-founder and longtime chief scientist Ilya Sutskever and CTO Mira Murati. OpenAI has also been hit with lawsuits from Alden Global Capital-owned newspapers alleging copyright infringement, as well as an injunction from Elon Musk to halt OpenAI’s transition to a for-profit.
In 2025, OpenAI is battling the perception that it’s ceding ground in the AI race to Chinese rivals like DeepSeek. The company has been trying to shore up its relationship with Washington as it simultaneously pursues an ambitious data center project, and as it reportedly lays the groundwork for one of the largest funding rounds in history.
Below, you’ll find a timeline of ChatGPT product updates and releases, starting with the latest, which we’ve been updating throughout the year. If you have any other questions, check out our ChatGPT FAQ here.
To see a list of 2024 updates, go here.
Timeline of the most recent ChatGPT updates
May 2025
OpenAI has launched a new feature for ChatGPT deep research to analyze code repositories on GitHub. The ChatGPT deep research feature is in beta and lets developers connect with GitHub to ask questions about codebases and engineering documents. The connector will soon be available for ChatGPT Plus, Pro, and Team users, with support for Enterprise and Education coming shortly, per an OpenAI spokesperson.
OpenAI launches a new data residency program in Asia
After introducing a data residency program in Europe in February, OpenAI has now launched a similar program in Asian countries including India, Japan, Singapore, and South Korea. The new program will be accessible to users of ChatGPT Enterprise, ChatGPT Edu, and API. It will help organizations in Asia meet their local data sovereignty requirements when using OpenAI’s products.
OpenAI to introduce a program to grow AI infrastructure
OpenAI is unveiling a program called OpenAI for Countries, which aims to develop the necessary local infrastructure to serve international AI clients better. The AI startup will work with governments to assist with increasing data center capacity and customizing OpenAI’s products to meet specific language and local needs. OpenAI for Countries is part of efforts to support the company’s expansion of its AI data center Project Stargate to new locations outside the U.S., per Bloomberg.
OpenAI promises to make changes to prevent future ChatGPT sycophancy
OpenAI has announced its plan to make changes to its procedures for updating the AI models that power ChatGPT, following an update that caused the platform to become overly sycophantic for many users.
April 2025
OpenAI clarifies the reason ChatGPT became overly flattering and agreeable
OpenAI has released a post on the recent sycophancy issues with the default AI model powering ChatGPT, GPT-4o, leading the company to revert an update to the model released last week. CEO Sam Altman acknowledged the issue on Sunday and confirmed two days later that the GPT-4o update was being rolled back. OpenAI is working on “additional fixes” to the model’s personality. Over the weekend, users on social media criticized the new model for making ChatGPT too validating and agreeable. It became a popular meme fast.
OpenAI is working to fix a “bug” that let minors engage in inappropriate conversations
An issue within OpenAI’s ChatGPT enabled the chatbot to create graphic erotic content for accounts registered by users under the age of 18, as demonstrated by TechCrunch’s testing, a fact later confirmed by OpenAI. “Protecting younger users is a top priority, and our Model Spec, which guides model behavior, clearly restricts sensitive content like erotica to narrow contexts such as scientific, historical, or news reporting,” a spokesperson told TechCrunch via email. “In this case, a bug allowed responses outside those guidelines, and we are actively deploying a fix to limit these generations.”
OpenAI has added a few features to its ChatGPT search, its web search tool in ChatGPT, to give users an improved online shopping experience. The company says people can ask super-specific questions using natural language and receive customized results. The chatbot provides recommendations, images, and reviews of products in various categories such as fashion, beauty, home goods, and electronics.
OpenAI wants its AI model to access cloud models for assistance
OpenAI leaders have been talking about allowing the open model to link up with OpenAI’s cloud-hosted models to improve its ability to respond to intricate questions, two sources familiar with the situation told TechCrunch.
OpenAI aims to make its new “open” AI model the best on the market
OpenAI is preparing to launch an AI system that will be openly accessible, allowing users to download it for free without any API restrictions. Aidan Clark, OpenAI’s VP of research, is spearheading the development of the open model, which is in the very early stages, sources familiar with the situation told TechCrunch.
OpenAI’s GPT-4.1 may be less aligned than earlier models
OpenAI released a new AI model called GPT-4.1 in mid-April. However, multiple independent tests indicate that the model is less reliable than previous OpenAI releases. The company skipped that step — sending safety cards for GPT-4.1 — claiming in a statement to TechCrunch that “GPT-4.1 is not a frontier model, so there won’t be a separate system card released for it.”
OpenAI’s o3 AI model scored lower than expected on a benchmark
Questions have been raised regarding OpenAI’s transparency and procedures for testing models after a difference in benchmark outcomes was detected by first- and third-party benchmark results for the o3 AI model. OpenAI introduced o3 in December, stating that the model could solve approximately 25% of questions on FrontierMath, a difficult math problem set. Epoch AI, the research institute behind FrontierMath, discovered that o3 achieved a score of approximately 10%, which was significantly lower than OpenAI’s top-reported score.
OpenAI unveils Flex processing for cheaper, slower AI tasks
OpenAI has launched a new API feature called Flex processing that allows users to use AI models at a lower cost but with slower response times and occasional resource unavailability. Flex processing is available in beta on the o3 and o4-mini reasoning models for non-production tasks like model evaluations, data enrichment, and asynchronous workloads.
OpenAI’s latest AI models now have a safeguard against biorisks
OpenAI has rolled out a new system to monitor its AI reasoning models, o3 and o4 mini, for biological and chemical threats. The system is designed to prevent models from giving advice that could potentially lead to harmful attacks, as stated in OpenAI’s safety report.
OpenAI launches its latest reasoning models, o3 and o4-mini
OpenAI has released two new reasoning models, o3 and o4 mini, just two days after launching GPT-4.1. The company claims o3 is the most advanced reasoning model it has developed, while o4-mini is said to provide a balance of price, speed, and performance. The new models stand out from previous reasoning models because they can use ChatGPT features like web browsing, coding, and image processing and generation. But they hallucinate more than several of OpenAI’s previous models.
OpenAI has added a new section to ChatGPT to offer easier access to AI-generated images for all user tiers
Open AI introduced a new section called “library” to make it easier for users to create images on mobile and web platforms, per the company’s X post.
OpenAI could “adjust” its safeguards if rivals release “high-risk” AI
OpenAI said on Tuesday that it might revise its safety standards if “another frontier AI developer releases a high-risk system without comparable safeguards.” The move shows how commercial AI developers face more pressure to rapidly implement models due to the increased competition.
OpenAI is currently in the early stages of developing its own social media platform to compete with Elon Musk’s X and Mark Zuckerberg’s Instagram and Threads, according to The Verge. It is unclear whether OpenAI intends to launch the social network as a standalone application or incorporate it into ChatGPT.
OpenAI will remove its largest AI model, GPT-4.5, from the API, in July
OpenAI will discontinue its largest AI model, GPT-4.5, from its API even though it was just launched in late February. GPT-4.5 will be available in a research preview for paying customers. Developers can use GPT-4.5 through OpenAI’s API until July 14; then, they will need to switch to GPT-4.1, which was released on April 14.
OpenAI unveils GPT-4.1 AI models that focus on coding capabilities
OpenAI has launched three members of the GPT-4.1 model — GPT-4.1, GPT-4.1 mini, and GPT-4.1 nano — with a specific focus on coding capabilities. It’s accessible via the OpenAI API but not ChatGPT. In the competition to develop advanced programming models, GPT-4.1 will rival AI models such as Google’s Gemini 2.5 Pro, Anthropic’s Claude 3.7 Sonnet, and DeepSeek’s upgraded V3.
OpenAI will discontinue ChatGPT’s GPT-4 at the end of April
OpenAI plans to sunset GPT-4, an AI model introduced more than two years ago, and replace it with GPT-4o, the current default model, per changelog. It will take effect on April 30. GPT-4 will remain available via OpenAI’s API.
OpenAI could release GPT-4.1 soon
OpenAI may launch several new AI models, including GPT-4.1, soon, The Verge reported, citing anonymous sources. GPT-4.1 would be an update of OpenAI’s GPT-4o, which was released last year. On the list of upcoming models are GPT-4.1 and smaller versions like GPT-4.1 mini and nano, per the report.
OpenAI started updating ChatGPT to enable the chatbot to remember previous conversations with a user and customize its responses based on that context. This feature is rolling out to ChatGPT Pro and Plus users first, excluding those in the U.K., EU, Iceland, Liechtenstein, Norway, and Switzerland.
OpenAI is working on watermarks for images made with ChatGPT
It looks like OpenAI is working on a watermarking feature for images generated using GPT-4o. AI researcher Tibor Blaho spotted a new “ImageGen” watermark feature in the new beta of ChatGPT’s Android app. Blaho also found mentions of other tools: “Structured Thoughts,” “Reasoning Recap,” “CoT Search Tool,” and “l1239dk1.”
OpenAI offers ChatGPT Plus for free to U.S., Canadian college students
OpenAI is offering its $20-per-month ChatGPT Plus subscription tier for free to all college students in the U.S. and Canada through the end of May. The offer will let millions of students use OpenAI’s premium service, which offers access to the company’s GPT-4o model, image generation, voice interaction, and research tools that are not available in the free version.
ChatGPT users have generated over 700M images so far
More than 130 million users have created over 700 million images since ChatGPT got the upgraded image generator on March 25, according to COO of OpenAI Brad Lightcap. The image generator was made available to all ChatGPT users on March 31, and went viral for being able to create Ghibli-style photos.
OpenAI’s o3 model could cost more to run than initial estimate
The Arc Prize Foundation, which develops the AI benchmark tool ARC-AGI, has updated the estimated computing costs for OpenAI’s o3 “reasoning” model managed by ARC-AGI. The organization originally estimated that the best-performing configuration of o3 it tested, o3 high, would cost approximately $3,000 to address a single problem. The Foundation now thinks the cost could be much higher, possibly around $30,000 per task.
OpenAI CEO says capacity issues will cause product delays
In a series of posts on X, OpenAI CEO Sam Altman said the company’s new image-generation tool’s popularity may cause product releases to be delayed. “We are getting things under control, but you 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,” he wrote.
March 2025
OpenAI plans to release a new ‘open’ AI language model
OpeanAI intends to release its “first” open language model since GPT-2 “in the coming months.” The company plans to host developer events to gather feedback and eventually showcase prototypes of the model. The first developer event is to be held in San Francisco, with sessions to follow in Europe and Asia.
OpenAI removes ChatGPT’s restrictions on image generation
OpenAI made a notable change to its content moderation policies after the success of its new image generator in ChatGPT, which went viral for being able to create Studio Ghibli-style images. The company has updated its policies to allow ChatGPT to generate images of public figures, hateful symbols, and racial features when requested. OpenAI had previously declined such prompts due to the potential controversy or harm they may cause. However, the company has now “evolved” its approach, as stated in a blog post published by Joanne Jang, the lead for OpenAI’s model behavior.
OpenAI adopts Anthropic’s standard for linking AI models with data
OpenAI wants to incorporate Anthropic’s Model Context Protocol (MCP) into all of its products, including the ChatGPT desktop app. MCP, an open-source standard, helps AI models generate more accurate and suitable responses to specific queries, and lets developers create bidirectional links between data sources and AI applications like chatbots. The protocol is currently available in the Agents SDK, and support for the ChatGPT desktop app and Responses API will be coming soon, OpenAI CEO Sam Altman said.
OpenAI’s viral Studio Ghibli-style images could raise AI copyright concerns
The latest update of the image generator on OpenAI’s ChatGPT has triggered a flood of AI-generated memes in the style of Studio Ghibli, the Japanese animation studio behind blockbuster films like “My Neighbor Totoro” and “Spirited Away.” The burgeoning mass of Ghibli-esque images have sparked concerns about whether OpenAI has violated copyright laws, especially since the company is already facing legal action for using source material without authorization.
OpenAI expects revenue to triple to $12.7 billion this year
OpenAI expects its revenue to triple to $12.7 billion in 2025, fueled by the performance of its paid AI software, Bloomberg reported, citing an anonymous source. While the startup doesn’t expect to reach positive cash flow until 2029, it expects revenue to increase significantly in 2026 to surpass $29.4 billion, the report said.
ChatGPT has upgraded its image-generation feature
OpenAI on Tuesday rolled out a major upgrade to ChatGPT’s image-generation capabilities: ChatGPT can now use the GPT-4o model to generate and edit images and photos directly. The feature went live earlier this week in ChatGPT and Sora, OpenAI’s AI video-generation tool, for subscribers of the company’s Pro plan, priced at $200 a month, and will be available soon to ChatGPT Plus subscribers and developers using the company’s API service. The company’s CEO Sam Altman said on Wednesday, however, that the release of the image generation feature to free users would be delayed due to higher demand than the company expected.
OpenAI announces leadership updates
Brad Lightcap, OpenAI’s chief operating officer, will lead the company’s global expansion and manage corporate partnerships as CEO Sam Altman shifts his focus to research and products, according to a blog post from OpenAI. Lightcap, who previously worked with Altman at Y Combinator, joined the Microsoft-backed startup in 2018. OpenAI also said Mark Chen would step into the expanded role of chief research officer, and Julia Villagra will take on the role of chief people officer.
OpenAI’s AI voice assistant now has advanced feature
OpenAI has updated its AI voice assistant with improved chatting capabilities, according to a video posted on Monday (March 24) to the company’s official media channels. The update enables real-time conversations, and the AI assistant is said to be more personable and interrupts users less often. Users on ChatGPT’s free tier can now access the new version of Advanced Voice Mode, while paying users will receive answers that are “more direct, engaging, concise, specific, and creative,” a spokesperson from OpenAI told TechCrunch.
OpenAI and Meta have separately engaged in discussions with Indian conglomerate Reliance Industries regarding potential collaborations to enhance their AI services in the country, per a report by The Information. One key topic being discussed is Reliance Jio distributing OpenAI’s ChatGPT. Reliance has proposed selling OpenAI’s models to businesses in India through an application programming interface (API) so they can incorporate AI into their operations. Meta also plans to bolster its presence in India by constructing a large 3GW data center in Jamnagar, Gujarat. OpenAI, Meta, and Reliance have not yet officially announced these plans.
OpenAI faces privacy complaint in Europe for chatbot’s defamatory hallucinations
Noyb, a privacy rights advocacy group, is supporting an individual in Norway who was shocked to discover that ChatGPT was providing false information about him, stating that he had been found guilty of killing two of his children and trying to harm the third. “The GDPR is clear. Personal data has to be accurate,” said Joakim Söderberg, data protection lawyer at Noyb, in a statement. “If it’s not, users have the right to have it changed to reflect the truth. Showing ChatGPT users a tiny disclaimer that the chatbot can make mistakes clearly isn’t enough. You can’t just spread false information and in the end add a small disclaimer saying that everything you said may just not be true.”
OpenAI upgrades its transcription and voice-generating AI models
OpenAI has added new transcription and voice-generating AI models to its APIs: a text-to-speech model, “gpt-4o-mini-tts,” that delivers more nuanced and realistic sounding speech, as well as two speech-to-text models called “gpt-4o-transcribe” and “gpt-4o-mini-transcribe”. The company claims they are improved versions of what was already there and that they hallucinate less.
OpenAI has launched o1-pro, a more powerful version of its o1
OpenAIhas introduced o1-pro in its developer API. OpenAI says its o1-pro uses more computing than its o1 “reasoning” AI model to deliver “consistently better responses.” It’s only accessible to select developers who have spent at least $5 on OpenAI API services. OpenAI charges $150 for every million tokens (about 750,000 words) input into the model and $600 for every million tokens the model produces. It costs twice as much as OpenAI’s GPT-4.5 for input and 10 times the price of regular o1.
OpenAI research lead Noam Brown thinks AI “reasoning” models could’ve arrived decades ago
Noam Brown, who heads AI reasoning research at OpenAI, thinks that certain types of AI models for “reasoning” could have been developed 20 years ago if researchers had understood the correct approach and algorithms.
OpenAI says it has trained an AI that’s “really good” at creative writing
OpenAI CEO Sam Altman said, in a post on X, that the company has trained a “new model” that’s “really good” at creative writing. He posted a lengthy sample from the model given the prompt “Please write a metafictional literary short story about AI and grief.” OpenAI has not extensively explored the use of AI for writing fiction. The company has mostly concentrated on challenges in rigid, predictable areas such as math and programming. And it turns out that it might not be that great at creative writing at all.
