Tag Archive from: openai

OpenAI recently introduced Operator, a smart AI agent that can independently perform tasks on the Internet. Operator goes beyond an ordinary chatbot and can visit websites, fill out forms and perform actions like a real user.

What is Operator and how does it work?

Operator is an AI system that can use a Web browser to execute commands. It works by visually analyzing Web pages and performing actions with a keyboard and mouse, just as a human would. This allows Operator to perform tasks such as reserving a table at a restaurant, ordering products from online stores and combining multiple actions simultaneously. For example, while Operator waits for confirmation, it can start searching for other options at the same time.

This AI system is based on a new model called Computer Using Agent (CUA), which is specially trained to use Web sites the way humans do: by looking, clicking and typing.

Operator in practice

During the launch, OpenAI demonstrated how Operator could independently reserve a table at a restaurant. In doing so, the AI system could even take into account location corrections and suggest other options when the desired time was not available.

It also showed how Operator can do online shopping for you. Based on an uploaded photo, the AI agent knew exactly which products to add and could complete the purchase independently. These are just a few examples of what you could use the AI agent for in everyday life.

Want to see this in action yourself? Watch Operator's demonstration video here.

Security and privacy

OpenAI has taken several measures to ensure user privacy and security. Operator always asks for permission before performing sensitive tasks, such as placing orders or logging into accounts. Users have the ability to take back control at any time.

To prevent abuse, Operator cannot perform malicious actions and automatically stops when a website is deemed suspicious. The system also ensures that user sessions remain completely private and no data is stored or shared without permission.

What will the future bring?

Operator is currently only available to Pro users in the United States, but will also be rolled out to Plus, Team and Enterprise users in the coming months. In the Netherlands, we still have to be patient, but it is clear that Operator will be of great impact.

As Operator continues to develop, it will get better and better at automating more difficult tasks and potentially supporting businesses in their day-to-day operations. This can help companies and marketers work even more efficiently and focus more on other important work.

The AI assistant of the future

The arrival of Operator offers a glimpse of a future where AI not only supports, but actually takes over tasks. For marketers, this means less time spent on routine tasks and more focus on strategic goals. By deploying this technology intelligently, companies can increase efficiency and strengthen their competitive position.

Take a leap forward in your marketing AI transformation every week

Every Friday, we bring you the latest insights, news and real-world examples on the impact of AI in the marketing world. Whether you want to improve your marketing efficiency, increase customer engagement, sharpen your marketing strategy or digitally transform your business. 'Marketing AI Friday' is your weekly guide.

Sign up for Marketing AI Friday for free.

Marketing AI Friday

OpenAI takes a big step toward full-fledged AI agents with the introduction of ChatGPT Tasks. This new beta feature, currently only available to paying users, allows you to schedule one-time and recurring tasks. This not only saves time, but also makes ChatGPT an even more effective assistant in marketing processes.

What is ChatGPT Tasks?

The Tasks feature allows users to set daily or weekly tasks that are automatically performed by ChatGPT. Examples include:

  • Weekly summaries of specific news within your field.
  • Daily personalized writing advice, such as for blogs, online courses or e-books.
  • Reminders for deadlines and other important appointments.
  • Customized exercise programs based on personal preferences and circumstances.

Based on previous conversations, ChatGPT can also proactively suggest tasks, making it increasingly responsive to user needs.

How do you set tasks?

Setting up Tasks is simple. You go through the following steps:

  1. From the menu, choose the 'ChatGPT-4o with scheduled tasks' model.
  2. In the chat, describe your desired task, including frequency and time.
  3. Manage your tasks via the Tasks section or directly in chat.
  4. Automatically receive results of your completed task via chat, email, web or mobile.

You can manage up to 10 active tasks simultaneously and divide them into specific projects for an organized workflow.

Constraints and future developments

Although the current features are promising, the beta version of ChatGPT Tasks still has some limitations. For example, it is not yet possible to link apps or files, and adjusting times can sometimes go wrong.

Yet this update appears to be just the beginning. Bloomberg previously reported that OpenAI is working on a more advanced AI agent called Operator, which can control computers independently. Work is also underway on Caterpillar, a feature that can perform deep research, explore Web sites and show results instantly in a sidebar. This would allow ChatGPT to be even better used in complex marketing research in the future.

How you can use ChatGPT Tasks

ChatGPT Tasks is currently only available to paid users. OpenAI has not yet announced if and when the feature will become available for free accounts.

Curious about how this new feature can support your marketing strategies? During our webinar "Marketing AI in Practice" on Friday, Jan. 31, we will go in-depth on the use and capabilities of this new ChatGPT feature. Don't miss it and learn how to automate your marketing tasks smarter.

Take a leap forward in your marketing AI transformation every week

Every Friday, we bring you the latest insights, news and real-world examples on the impact of AI in the marketing world. Whether you want to improve your marketing efficiency, increase customer engagement, sharpen your marketing strategy or digitally transform your business. 'Marketing AI Friday' is your weekly guide.

Sign up for Marketing AI Friday for free.

Marketing AI Friday

In March 2024, Sam Altman, CEO of OpenAI, indicated in an interview with Lex Fridman that he was not enthusiastic about building a search engine like Google. However, he did see opportunities in combining AI with online search, something no one has yet fully realized. On Oct. 31, 2024, OpenAI looks to make that vision a reality with the launch of ChatGPT Search. The question is: Will this cause us to search in a different way?

ChatGPT Search: How does it work?

ChatGPT Search is a new feature that allows users to find information directly online, without the familiar Google ads and long lists of search results. Where Google often displays an extensive list of clickable links, ChatGPT Search provides a clear answer immediately. This ensures that you get the information you are looking for faster and without distraction, plain and simple.

ChatGPT Search's capabilities go beyond text. For example, the platform supports up-to-date information such as stock quotes, complete with interactive charts that can be viewed in detail. The new Chrome extension also allows users to enter search queries directly into the browser, after which ChatGPT Search provides immediate answers in a convenient overlay window. This seamless interaction makes search not only faster, but also more accessible for everyday use.

What does ChatGPT Search mean for SEO?

ChatGPT Search operates without ads, providing a different experience than search engines such as Google, Bing and Yahoo, which rely largely on ad revenue. Google, for example, derives 75% of its revenue from ads. With the rise of AI-driven search methods, the demand for alternative ways of searching is growing, which can affect how companies are visible to their target audience. This requires a customized strategy to continue reaching customers effectively.

Companies that previously relied on Google traffic can use ChatGPT Search to rely on visitors who are already better informed before they arrive at the website. This change is likely to affect not only the quantity but also the quality of traffic to websites. The targeting of AI search engines like ChatGPT Search provides users with answers based on their specific questions, meaning that visitors typically have more background information about your company or services before they even reach your website.

AIO: The key to visibility in AI search engines

The rise of AI search engines such as ChatGPT, Claude, Perplexity and Gemini presents new opportunities and challenges for content marketing and SEO. Whereas traditional SEO focuses primarily on findability within Google, AIO (AI-optimized) content specifically targets AI platforms. AIO goes beyond keyword optimization; it is all about context and relevance so that companies are not only visible, but also maintain their visibility within the AI search environment.

At NXTLI, we are exploring the potential of AIO strategies to better align content with AI search engines. In addition to focusing on Google, we are working on ways to develop content strategies that can also be effective within emerging AI search channels. Want to learn more about AIO? Then read our article on it.

Google's position in an AI-driven world

Google has some significant advantages with its vast dataset and market dominance, but the company faces the challenge of staying competitive. Google is likely to increase its focus on AI-driven and personalized search results to strengthen their current position. Nevertheless, it remains uncertain whether this will be enough to surpass the success of alternatives such as ChatGPT Search.

Google has a strong position, and many people will continue to use Google out of habit. Yet the demand for ad-free and user-friendly search options continues to grow, increasing the pressure on Google to innovate its technology. Meanwhile, OpenAI, with its alternative approach, could potentially gain a stronger market position and become a serious competitor to Google.

What does ChatGPT Search mean for businesses?

It is more important than ever to write content that truly connects with what the searcher wants to know or find. The rise of new AI search engines such as ChatGPT Search requires an approach in which content is clear, relevant and valuable. While Google still plays an important role, strong content strengthens your position on all platforms and ensures that you are providing the right answers to your target audience's questions wherever they search. Need help with that and want to spar with us to hone your content strategy? Then get in touch with us.

Take a leap forward in your marketing AI transformation every week

Every Friday, we bring you the latest insights, news and real-world examples on the impact of AI in the marketing world. Whether you want to improve your marketing efficiency, increase customer engagement, sharpen your marketing strategy or digitally transform your business. 'Marketing AI Friday' is your weekly guide.

Sign up for Marketing AI Friday for free.

Marketing AI Friday

ChatGPT is one of the most widely used AI chatbots worldwide. Because there are many different versions, it can be difficult to determine which one best suits your specific needs. Should you go for the standard version or a more advanced one? In this article, we will take a look at the different versions of ChatGPT and help you choose the right one for your purpose.

1. ChatGPT-4o

ChatGPT-4o is the standard version suitable for most tasks. It is ideal for everyday use, such as writing texts, translations, summarizing information and even simple programming tasks. This version also supports the ability to upload files and has access to the Internet, which makes it extra versatile. If you are looking for a reliable version of ChatGPT, ChatGPT-4o is a good choice.

Ideal for:

  • Writing blogs and articles
  • Translation of texts
  • Coding and debugging
  • General Q&A

ChatGPT-4o has been available since May 13, 2024.

Read more about ChatGPT-4o here.

2. ChatGPT-4o Canvas

Are you working on a project that requires frequent changes to code or text? Then the version with Canvas is perfect for you. With features like version control and a handy work screen, you can easily edit text and code without losing previous versions. This makes working with ChatGPT much more efficient, especially if you're doing a lot of writing or code.

Ideal for:

  • Version control in complex projects
  • Real-time collaboration on documents and code
  • Quick adjustments and enhancements without losing previous versions

ChatGPT-4o Canvas has been available since Oct. 3, 2024.

Read more about ChatGPT-4o with Canvas here.

3. o1

Are you working on projects that require in-depth analysis and creativity? Then o1 is a great tool. This version is optimized for incremental reasoning and understands complex instructions.

Ideal for:

  • Advanced analytical tasks
  • Creative brainstorming sessions
  • Projects requiring complex problem solving

o1 has been available since December 5, 2024.

Read more about o1 here.

4. o1-mini: Quick answers to simple questions

If you are looking for speed and efficiency for simple tasks, o1-mini is the right choice. This lightweight version offers faster responses without sacrificing quality. Perfect for those who need quick answers to simple questions or tasks without a lot of computing power.

Ideal for:

  • Quick answers to simple questions
  • Simple prompts and commands
  • Tasks that require little processing power

o1-mini has been available since December 5, 2024.

Read more about o1-mini here.

5. ChatGPT Pro: most advanced subscription

This subscription, of $200 per month, gives unlimited access to o1, o1-mini, GPT-4o and advanced voice features.

The o1 Pro version uses more computing power for complex questions and consistently scores better in tests on data science, programming and legal analysis. In particular, the reliability of answers is a lot better than the standard version on complex mathematical, technical and scientific questions.

Ideal for:

  • Professionals who must solve complex problems on a daily basis, such as researchers and engineers.
  • Really challenging technical or analytical issues
  • Tasks that require a lot of computing power

ChatGPT Pro has been available since December 5, 2024.

Read more about ChatGPT Pro here.

Which version best suits your needs?

Briefly:

  • Most standard tasks: ChatGPT-4o
  • Extended elaboration of text or code: ChatGPT 4o Canvas
  • Quick answers to simple questions: o1-mini
  • Complex tasks requiring advanced step-by-step "reasoning": o1

By choosing the right version of ChatGPT, you can take full advantage of AI technology and make your work more efficient and effective. Which version best suits your needs?

Take a leap forward in your marketing AI transformation every week

Every Friday, we bring you the latest insights, news and real-world examples on the impact of AI in the marketing world. Whether you want to improve your marketing efficiency, increase customer engagement, sharpen your marketing strategy or digitally transform your business. 'Marketing AI Friday' is your weekly guide.

Sign up for Marketing AI Friday for free.

Marketing AI Friday

Last week was a busy one for OpenAI. In addition to raising a whopping $6.6 billion in funding, bringing the company to an impressive $157 billion valuation, CEO Sam Altman also had time for DevDay, the annual event where OpenAI presents the latest technologies and innovations to developers.

Several exciting announcements were made during DevDay that could change the future of AI. Below is an overview of the major updates and new features that OpenAI unveiled:

Direct interactions with AI through the real-time API

A key announcement is the new Realtime API, which allows developers to build AI-driven apps that function like real conversation partners. Through constant connection, the AI understands nuances such as emotions and accents, making for a more natural conversation. Another powerful part of this API is the ability to use function calls, which allow the AI to not only respond to queries, but also perform actions within the app. For example, consider automatically placing an order or performing a task based on a spoken command.

Speed and cost savings thanks to prompt caching

Another innovation is Prompt Caching, a feature that remembers and reuses previous AI tasks. This is especially useful if you regularly perform similar tasks, such as text generation or analysis. Prompt Caching makes execution up to 50% cheaper and faster. It works with all new OpenAI models. This can be especially interesting for companies that work with AI on a large scale and want to save costs.

Train smart and efficient AIs with model distillation

Model Distillation allows you to train smaller, specialized AI models with the knowledge of larger models. This allows you to create, for example, an AI that knows exactly how your brand communicates, without having to use the large and expensive models. This produces faster and more efficient results, making it attractive for companies to develop custom AI solutions.

Make AI understand images with vision fine-tuning

With the introduction of Vision Fine-Tuning, GPT-4 becomes even more powerful. The model can now also understand images, enabling numerous new applications, such as smart classification of products or improving image recognition. Developers can now train the model with both text and images, significantly increasing the potential for visual AI applications.

New Features in the Playground

For developers who like to experiment with AI, OpenAI has also added new features to the Playground. The new "Generate" button provides support for creating system prompts, function definitions and output schedules. This makes it easier to set up complex tasks quickly and efficiently.

Reduced costs for APIs

An added bonus for developers is that working with GPT-4 via the API is now cheaper. This lowers the barrier for companies and individuals to integrate advanced AI into their workflows. Also, OpenAI has announced that prices for the Realtime API for voice (now around €0.06 per minute of input and €0.23 per minute of output) are likely to drop soon, making the use of voice-driven AI apps even more attractive.

Future of voice-activated AI apps

With these announcements, OpenAI is taking a clear step toward a future in which real-time, voice-driven AI apps become the norm. The new features make it easier AND cheaper for developers to build AI solutions that can interact with users in real time. This opens the door to a wave of innovative applications across industries, from customer service to productivity tools.

OpenAI has indicated that developers can now try out these new features for free with the training tokens that are temporarily available. This provides a great opportunity to experiment and see how these new tools can be used for your projects.

Take a leap forward in your marketing AI transformation every week

Every Friday, we bring you the latest insights, news and real-world examples on the impact of AI in the marketing world. Whether you want to improve your marketing efficiency, increase customer engagement, sharpen your marketing strategy or digitally transform your business. 'Marketing AI Friday' is your weekly guide.

Sign up for Marketing AI Friday for free.

Marketing AI Friday

OpenAI recently began rolling out the long-awaited "Advanced Voice Mode" within their ChatGPT app, where users can take advantage of a more human interaction with artificial intelligence. This technology uses advanced speech recognition and voice production, making conversations with AI feel more natural than ever. However, the rollout in Europe is still challenging and not all users have access to these new features.

While some users who are within a Teams subscription can already try out the new voice feature, others report that the feature is not yet available to them. Online sources indicate that the rollout is phased, with not all users in Europe getting access at the same time. This leads to confusion and ambiguity, especially since OpenAI has not given an exact timeline for full availability in Europe.

What makes Advanced Voice Mode different?

ChatGPT's advanced voice feature is not simply an upgrade to existing speech technology. While many other AI models, such as those from Google and Meta, focus on text-to-speech features, OpenAI goes a step further by offering multimodal AI. This means that the technology not only converts words into speech, but can also analyze your emotions and tone, making the interaction more human. With this development, ChatGPT allows users to communicate as if they were talking to a real person, capturing and processing nuances in speech and emotion.

The challenges of regulation in Europe

Despite the impressive technology, the European rollout of Advanced Voice Mode is running into obstacles. OpenAI itself has not given a specific reason for the delay, but it is likely that strict AI regulations in Europe are playing a major role in this. The European Union is currently working to implement the AI Act, which sets strict rules for the use of artificial intelligence, especially in the areas of privacy and ethics. Because ChatGPT's Advanced Voice Mode potentially processes sensitive data, such as voice recordings and emotional expressions, this feature may not yet fully comply with the requirements of the AI Act.

In an effort to address the issues, OpenAI recently announced that they have signed on to the core obligations of the EU AI Pact . This is a step in the right direction of full compliance with European legislation. Still, it remains unclear when Advanced Voice Mode will be fully rolled out in the EU for all users.

Inconsistent experiences among users

Meanwhile, users inside and outside Europe report very different experiences with the new voice feature. While some Plus or Teams subscription users are able to use Advanced Voice Mode right away, others, with similar subscriptions, report that they have yet to access the feature. This phased rollout is confusing, especially since OpenAI has not shared any specific information about the order of rollout.

Our experience

We were able to test the new speech feature and our experiences are mixed. On the plus side, conversations feel much more natural, with a markedly better understanding of context and nuances. The ability to pause conversations without confusing the AI is a big improvement. Even the AI's singing abilities add a humorous element.

Still, there are some challenges. For example, conversations are sometimes interrupted unexpectedly and the AI does not always understand questions correctly. In addition, the chosen voice suddenly changes and the voice feature does not yet work with custom GPTs. Finally, it does not yet have Internet or file upload access.

The impact on AI

OpenAI's introduction of Advanced Voice Mode is an important step in the development of voice technology. While Google and Meta are working on their own versions of multimodal voice models, OpenAI is one of the first to make this technology widely available. Businesses and individuals can benefit from more natural interactions with their customers, which can ultimately lead to a better experience when using ChatGPT.

One of the most striking features of Advanced Voice Mode is how it is able to recognize emotions and speech patterns. This makes it possible to use AI in a more human way, especially where showing empathy is important.

Take a leap forward in your marketing AI transformation every week

Every Friday, we bring you the latest insights, news and real-world examples on the impact of AI in the marketing world. Whether you want to improve your marketing efficiency, increase customer engagement, sharpen your marketing strategy or digitally transform your business. 'Marketing AI Friday' is your weekly guide.

Sign up for Marketing AI Friday for free.

Marketing AI Friday

The AI world is eagerly awaiting the introduction of Orion, OpenAI's latest project expected to launch in early 2025. Orion, the successor to GPT-4, brings significant improvements in reasoning and multimodal processing. It promises to be an important step in the development of artificial intelligence. Many speculate about the impact this model will have on various industries.

What makes Orion different?

Orion stands out for its focus on both reasoning and multimodal processing, meaning that it can not only process text, but also understand and analyze images and videos. This enables it to perform complex tasks such as image recognition, video reviews, and simultaneous analysis of different types of media content. With this, Orion goes far beyond GPT-4, which focused primarily on text processing.

The introduction of multimodal processing could revolutionize how AI is deployed within different industries. Consider, for example, healthcare, where AI can analyze images such as X-rays and scans in conjunction with medical records, or the entertainment industry, where AI can edit or interpret videos for different purposes. This versatility makes Orion one of the most promising AI developments of the moment.

Orion's impact on the AI industry

Orion has the potential to change the way companies deploy AI. Its ability to analyze multiple types of data such as text, image and video simultaneously gives companies an edge in optimizing their processes and gaining deeper insights into complex issues.

For example, marketing teams can improve their campaigns by automatically analyzing which visual content performs best with different audiences. In healthcare, Orion can help accelerate and refine diagnoses by processing medical images and records simultaneously. Even in education, Orion can provide more personalized learning experiences by reviewing both textual and visual content and responding effectively.

The future of Orion

While the full unveiling of Orion is still some time away, it is clear that this development has the potential to have a major impact on multiple industries. OpenAI continues to push the boundaries of what is possible with AI, and with the arrival of Orion, we are one step closer to a future where AI not only takes over human tasks, but performs them smarter and more efficiently.

For companies looking to move forward with AI, Orion is a development to keep an eye on. Its ability to accelerate processes, improve insights and drive better results will be hugely transformative for industries such as marketing, healthcare and education.

Take a leap forward in your marketing AI transformation every week

Every Friday, we bring you the latest insights, news and real-world examples on the impact of AI in the marketing world. Whether you want to improve your marketing efficiency, increase customer engagement, sharpen your marketing strategy or digitally transform your business. 'Marketing AI Friday' is your weekly guide.

Sign up for Marketing AI Friday for free.

Marketing AI Friday

OpenAI launched Sept. 12, 2024, the first version of their new 'Strawberry' series of AI models: ChatGPT-o1. This makes it the first company to move from AI Level 1 to Level 2. The name o1 can be seen as a reset (from the 'old' GPT-4 model of which Sam Altman previously stated 'it sucks') to the beginning of a new phase. 

What levels of AI are there?

According to OpenAI's vision of the development toward artificial general intelligence (AGI), this happens in 5 different stages:

  1. Conversational AI (such as ChatGPT-4 and current models of Claude, Gemini, Grok and Llama): Conducts natural conversations.
  2. Reasoning AI (ChatGPT-o1): Solves basic problems similar to someone at the level of a PhD student or who has earned a doctorate. AI outperforms 50% of humans on the tasks they are skilled at.
  3. Autonomous AI: Acts independently for extended periods of time without human intervention. AI outperforms 90% of humans on the tasks they are skilled at.
  4. Innovating AI: Develops new ideas and solutions. AI outperforms 99% of humans on the tasks they are skilled at.
  5. Organizational AI: The ultimate level of AI researchers where it performs tasks of an entire organization and outperforms 100% of its people. 

OpenAI expects to reach level 5 (AGI) around 2030, a little sooner or a little later!

agi steps Bloomberg

The difference between GPT-o1 and GPT-4o

What makes GPT-o1 different from GPT-4o? Whereas with ChatGPT-4o (level 1) you get an answer almost immediately, these new level 2 models can think longer and incrementally to give better answers and follow instructions better. With the answer, you can also see what steps it followed and how long it took. Here you can see ChatGPT-o1 in action. I asked the infamous question "how many r's are in the word 'strawberry'? You can see how ChatGPT-4o and o1 handle this simple-seeming question that all AI models have so far failed to address.

chatgpto1

The performance of ChatGPT-o1

In various tests, ChatGPT-o1 showed impressive results: it solved 83% of the problems in the International Mathematics Olympiad qualifier, while GPT-4o achieved only 13%. In coding benchmarks, o1 outperformed 89% of participants in Codeforces competitions.

International Mathematics Olympiad qualifier ChatGPTo1

The o1 models use advanced reinforcement learning techniques. This allows them to refine their thinking process, try different strategies and even learn from their mistakes. OpenAI has also made great strides in the area of security. The o1 preview model scored 84 on one of OpenAI's most difficult jailbreaking tests, compared with just 22 for GPT-4o.  

Getting started with ChatGPT-o1

If you have a Plus or Teams subscription, you can use ChatGPT-o1 from now on by selecting 'o1 preview' or 'o1 mini' in ChatGPT at the top.

Try o1 especially for complex tasks or multi-step problems to experience the differences and get an idea of where this development is going. In this first preview version, several features of ChatGPT-4o are still missing, such as the voice feature, accessing the Internet, invoking GPTs and uploading files and images.

OpenAI indicates that these features will be added soon. For now, it is still best to use the ChatGPT-4o version for the tasks where you need it. With the introduction of these "self-thinking" AIs, we are entering a new phase where we can use AI for more complex tasks and the outcomes will become much better. This gives us very interesting opportunities, but also challenges to adapt our role and way of working to this. 

Invitation to monthly AI inspiration sessions

We would love to take a look at the most important AI developments with you, discuss the latest news and updates, share experiences, give you concrete tools, you can ask questions and spar with us.

Do you enjoy being a part of this as well? Then sign up.

Banner Online AI inspiration session

The world of artificial intelligence is not standing still. While many companies are just starting to use AI, OpenAI already seems to be ready for the next big step. Rumors abound about what their next development will bring, and more and more eyes are on an upcoming release Strawberry.

According to a recent article by The Information, OpenAI is working on a new GPT version that builds on previous models, focusing on better problem-solving skills and better reasoning. If they succeed in this, it could give them another big lead in the race for the most advanced AI technology.

What can we expect from Strawberry?

Strawberry is seen as a big step forward from current AI models, such as GPT-4. In particular, it is said to be better at solving complex mathematical problems, providing assistance in strategic planning and even in tasks that require human logic, such as solving puzzles. These improvements focus on deeper and more accurate reasoning, which is essential for more challenging tasks.

When is Strawberry coming out?

Although the full release of Strawberry is probably not expected until 2025, a beta version will be available as early as this fall. This will give users and companies a chance to experiment with the model in advance, while OpenAI continues to work on refining the technology.

Update: OpenAI launched ChatGPT-o1, which this article is about, on Sept. 12, 2024.

What makes Strawberry different from GPT-4?

One notable difference from earlier models is that OpenAI's Strawberry is given more time to arrive at answers. Whereas GPT-4 and other similar AIs provide quick answers, Strawberry is designed to think more deeply about complex problems. This means that the model will be able to provide more accurate and reasoned answers to complex questions.

OpenAI and responsible use of AI

In the quest for increasingly sophisticated AI models, OpenAI is also keeping an eye on safety and accountability. CEO Sam Altman recently announced a partnership with the US AI Safety Institute to ensure that new AI models, such as Strawberry, are thoroughly tested before widespread deployment. This demonstrates the importance of careful development and responsible use of AI.

What is the impact of this new generation of AI?

While it is not yet clear how big the impact of Strawberry will be, it is clear that new AI systems such as these will continue to change the way we work. They offer promising opportunities for innovation and efficiency, but they will also affect roles and tasks within organizations. It is therefore wise to keep a close eye on these developments and consider how your company can prepare for these changes.

Looking ahead to Orion

In addition to Strawberry, OpenAI is also developing a new AI model called Orion. This project should ensure that OpenAI is even further ahead. Interestingly, Strawberry is playing an important role in the development of Orion by generating the training data needed to improve this model.

The future of AI is brimming with possibilities, and with releases like Strawberry and Orion, OpenAI shows that they continue to innovate. For companies and users following this technology, this could mean another big step in how AI impacts our daily lives.

Take a leap forward in your marketing AI transformation every week

Every Friday, we bring you the latest insights, news and real-world examples on the impact of AI in the marketing world. Whether you want to improve your marketing efficiency, increase customer engagement, sharpen your marketing strategy or digitally transform your business. 'Marketing AI Friday' is your weekly guide.

Sign up for Marketing AI Friday for free.

Marketing AI Friday

In August 2024, OpenAI once again gained attention by launching the latest update to its GPT-4o model. This update, released on Aug. 8 brought the company back to the top of the AI rankings, tested by Chatbot Arena, after being briefly overtaken by Gemini 1.5.

What makes this return special is that OpenAI did not introduce an entirely new model, but rather thoroughly improved the existing GPT-4o model.

ChatGPT-4o is now ranked No. 1 in the following categories:

  • General
  • Mathematics
  • Coding
  • Complex prompts
  • Instructional follow-up
  • Longer questions
  • Multi-step interviews

Improved performance and better problem solving

The focus of the GPT-4o update was primarily on optimizing performance and addressing existing bugs.

Instead of choosing to launch an entirely new model, OpenAI chose to build on the strong foundation of GPT-4o and refine it further. They did this by thoroughly purifying the dataset, replacing outdated or inaccurate information with current and accurate data. As a result, the model has now taken a significant leap forward, both in speed and efficiency.

Complex prompts are answered better

One notable improvement of the updated GPT-4o model is how well it now handles complex prompts. While previous versions sometimes had trouble understanding complex prompts, with this update it seems to be very accurate and fast at handling complex issues. This makes the model especially useful for professionals and companies that use AI for tasks that require a high level of precision. These include analyzing complex data sets or developing sophisticated software.

Better instructional follow-up

Furthermore, GPT-4o is now also better able to accurately follow long and detailed instructions. This was an issue where previous versions of the model sometimes fell short, resulting in output errors. The update has addressed these shortcomings, so users can now rely on an AI that executes instructions more efficiently and consistently. This is especially valuable for companies using AI to automate day-to-day tasks and complex workflows.

Growing impact of GPT-4o

The improvements in GPT-4o not only strengthen OpenAI's position, but also provide tangible benefits to users worldwide. The model is now even more capable of performing a wide range of tasks with a higher degree of reliability and accuracy.

Signal to competitors

This update demonstrates the importance of continuous innovation within AI. Although Google's Gemini 1.5 temporarily took the lead, OpenAI's rapid response shows that continuous improvement of existing models is just as important as developing new ones.

Invitation to monthly AI inspiration sessions

Together with you, we would like to take a look at the most important AI developments, discuss the latest news and updates, share experiences, inspire each other, take advantage of what is possible and prepare ourselves as well as possible for the changes that are coming.

Would you like to attend? Then sign up.

Banner Online AI inspiration session