Latest news with #O3
Time of India
08-08-2025
- Time of India
OpenAI launches ChatGPT-5 and the new version might just replace your therapist, tutor, and best friend, here's why
Say hello to the latest AI brainchild from OpenAI, ChatGPT-5. After much anticipation, Sam Altman's team has finally pulled back the curtain, and spoiler alert: it is not just a glow-up; it is a full-blown AI revolution. If GPT-4 was smart, GPT-5 is valedictorian, class president, and prom queen, all at once. ChatGPT-5 is officially here According to OpenAI insiders, GPT-5 is not just one model but a unified beast. It blends quick-answering efficiency with deep, thoughtful reasoning, and includes a smart router that decides in real-time which thinking mode to activate depending on your query's vibe. And yes, if you literally type 'think hard about this,' it listens. Who gets access and what's the catch? Everyone, well, sort of. ChatGPT-5 is now rolling out to Free, Plus, Pro, Team, and Enterprise users. However, there is a twist: Free users get it with a usage limit. Once that runs out, the system slides you into GPT-5 Mini mode, still smart, but a little less brainy. For paying users (bless your wallet), GPT-5 is the default experience. Pro users even get a turbocharged version. Team, Enterprise, and Edu users can all make GPT-5 their go-to assistant for daily grind, student help, or mega-corp brainstorming sessions. GPT-5 features you need to know about This is not your average chatbot. GPT-5 has been upgraded to be significantly faster, sharper, and way more aligned with real-world needs. It knows when to give a snappy reply and when to dive deep with a dissertation-level answer. Whether you are crafting essays, coding mini-games, or decoding health info, GPT-5 is on it. One insider even spilled that GPT-5 can now produce games like rolling ball simulators or lo-fi visualisers from just a single prompt. Coding girlies and guys, your AI BFF has arrived. And yes, hallucinations (aka those rogue, made-up facts) have been dramatically reduced. Instruction-following? Sharper than ever. Basically, it has fewer oops moments and more 'wow, how did you know that?' vibes. GPT-5 vs GPT-4o comparison Let us be real: GPT-4o had its moment, but GPT-5 is eating it for breakfast. OpenAI claims GPT-5 responses are 45 percent less likely to be factually wrong. Compared to OpenAI's O3 model, that number jumps to 80 percent. It is like comparing a calculator to a full-blown genius. GPT-5 also takes over the duties of older models like GPT-4.1, GPT-4.5, OpenAI o3 and o4-mini. Basically, it is the new standard, the main character, the era. The folks at OpenAI are planning to make this unified model even more seamless. That means no more thinking about switching models — the system will smartly handle it all for you. The router improves over time, based on what you like, switch to, or even rate better. Creepy? A little. Useful? Absolutely. ChatGPT-5 is not just a tech upgrade, it is a vibe shift. Whether you are a student trying to survive finals, a developer coding like a wizard, or just someone who wants a witty AI buddy, this model is your new best friend. It is clever, fast, and finally feels like the AI sidekick we have all been promised in sci-fi movies.
Hindustan Times
06-08-2025
- Business
- Hindustan Times
OpenAI's latest GPT-OSS AI models can run on your phone and laptop: What's new
Ahead of the highly anticipated GPT-5 AI model launch, the successor to GPT-4, OpenAI has launched two new open-weight AI models called gpt-oss120b and gpt-oss-20b. The AI giant claims that these will be available under the flexible Apache 2.0 licence. This essentially allows you to download and customise it yourself, and then eventually run it on a device as small as a laptop. They are available to download using Hugging Face, which means they are freely accessible to developers. OpenAI CEO Sam Altman claims that these new models are a "big deal."(AFP File) OpenAI says that these are optimised for deployment on consumer-grade hardware, such as laptops, and are trained using a mix of reinforcement learning and techniques informed by OpenAI's more advanced models like O3. OpenAI also claims that the gpt-oss120b model is nearly on par with OpenAI's O4 mini for core reasoning benchmarks while running on an 80 GB GPU. On the other hand, gpt-oss-20b delivers results similar to OpenAI's O3 mini on common benchmarks while running on edge devices with just 16 GB of GPU memory. What can they do? It is worth noting that these models are not multimodal, so they cannot generate items like images or videos, they can only generate pure text. Furthermore, OpenAI says that these are designed to be used within agentic workflows with tools like web search or Python code execution, and they possess reasoning capabilities, which include the ability to adjust the reasoning effort. OpenAI says that they are also customisable, provide a full chain of thought, and support structured outputs. Sam Altman calls new models a 'big deal' Sam Altman, the OpenAI CEO, called it a "big deal", taking to X to post about the new release. He said that GPT-OSS is a state-of-the-art, open weights reasoning model with strong real-world performance comparable to O4 mini, and is something that you can run locally on your computer or even phone. Altman also claimed that OpenAI believes this is the best and most usable open model in the world. He added that OpenAI believes more good than bad will come from it and that the company has worked hard to mitigate more serious safety issues, especially around biosecurity. MOBILE FINDER: iPhone 16 LATEST Price
International Business Times
30-07-2025
- Business
- International Business Times
OpenAI's Powerful GPT-5 Model May Launch Next Week, Promises Major Upgrades
OpenAI is likely to unveil its highly anticipated GPT-5 model as early as next week. The new model is expected to bring major improvements in reasoning, memory, and real-world task automation. Industry insiders say it could be the most advanced model yet, combining the strengths of OpenAI's earlier O3 and O4 models. CEO Sam Altman has hinted at the release, even showcasing GPT-5's abilities in a podcast interview. He admitted feeling "useless" after the model answered a question he couldn't. Reports claim GPT-5 has already been tested by early users and may launch with smaller "mini" and "nano" versions for specific uses. Smarter, Unified, and Multimodal GPT-5 is expected to offer a unified experience by blending features of multiple GPT versions. This integration is likely to enhance performance in logic, problem-solving, and coding. The model may display near PhD-level reasoning, as per initial tests. It will be available via ChatGPT and OpenAI's API. The AI is also rumored to support enhanced multimodal interaction, including video alongside text, voice, and images. This aims to make user interactions smoother and more natural. The new model could also include better memory, potentially handling over 256,000 tokens — double that of GPT-4o. Towards Autonomous AI Agents Another big leap may be in autonomy. GPT-5 is said to handle complex tasks independently. It could manage digital tools and complete multi-step online tasks with minimal human help, acting more like a virtual assistant. These features mark a big step toward OpenAI's goal of achieving artificial general intelligence. While OpenAI hasn't officially confirmed all these features, expectations are high for what could be a landmark release in AI technology.
International Business Times
17-07-2025
- Business
- International Business Times
Who are Jason Wei and Hyung Won Chung? Meta Hires 2 More OpenAI Engineers
Meta, the parent of Facebook, has recruited two high-profile researchers from OpenAI—Jason Wei and Hyung Won Chung—in its latest AI hiring spree. The move is the latest sign of Meta's increasing ambition to chart the future of artificial intelligence and contend with rivals like Google DeepMind and Apple. X Wired reported that OpenAI has already disabled Wei and Chung's internal Slack accounts to verify their departure. The two researchers are said to have collaborated on many of the big OpenAI projects before moving to Meta, including deep search and large language models O1 and O3. Jason Wei, who joined OpenAI in 2023, was formerly employed by Google. He has spent the past 40 years mastering chain-of-thought reasoning in AI, which is the process of training AI to solve complex problems one step at a time. Wei also likes reinforcement learning—a technique in which AI models are rewarded for making correct choices, which is now pivotal to the AI revolution. Hyung Won Chung, who also joined OpenAI in 2023, was a collaborator of Wei's on similar projects. Much of his research is geared toward logical aspects of AI, in particular for reasoning systems and agents. The duo had also been colleagues at Google, so their move to Meta was a notable double hire. Meta's recent hiring rush has pulled in top AI talent from some of the biggest names in the technology sector. Some recruits to its AI division, most notably those working on superintelligence projects, have been offered as much as $300 million over four years, according to reports. But not everyone is appreciating this talent grab from the rival organizations. Dell Technologies CEO Michael Dell expressed concerns about the cultural impacts of such fierce recruitment efforts. In a podcast with venture capitalists Bill Gurley and Brad Gerstner, Dell said that high salaries for even new employees could elicit resentment from long-tenured coworkers. "It's going to be a cultural challenge, no doubt," Dell said. He cautioned that Meta's internal teams may become divided if current employees feel ignored or underappreciated in the face of the huge pay packages that newcomers are receiving.
Hans India
06-07-2025
- Business
- Hans India
Why responsible AI is key for building better, brighter future for the world
Artificial intelligence has become the defining technology of our era, with recent years marking remarkable milestones in AI development. As the debate around AI technology heats up, people are actively discussing its usage, impact, and transformative potential across industries. However, amidst all the excitement and concern, there's notably less conversation about the responsibility that comes with this powerful technology. This gap in discourse presents both a challenge and an opportunity to shape how we approach AI governance. Breakthrough models like OpenAI's O3 achieved near-human performance on complex reasoning tasks, while AI applications continue to expand across industries in ways that seemed impossible just a few years ago. Yet as we witness this extraordinary progress, we find ourselves at a crucial juncture where thoughtful governance can shape AI's trajectory for maximum human benefit. The question isn't whether AI will transform our world - it's how we'll guide that transformation. Historically, humanity has taken a reactive approach to governing powerful technologies. The atomic bombing and destruction of two cities during the World War II ultimately paved the way for the formation of the United Nations - a response born from tragedy rather than foresight. Rather than repeating this pattern of waiting for catastrophe to drive action, we have an unprecedented opportunity with AI to be proactive. This is the right time for us to debate and establish a world governing council for responsible AI, creating international governance frameworks while embracing AI's transformative potential. This proactive approach represents a fundamental shift from reactive crisis management to thoughtful stewardship of emerging technology. Frankly speaking, recent developments highlight both AI's transformative power and the importance of responsible deployment. Google's Gemini 2.0 and Anthropic's Claude 4 demonstrate unprecedented capabilities in autonomous planning and execution, opening new possibilities for scientific research and problem-solving. However, this rapid advancement brings important considerations that deserve our attention. Workforce transitions affecting 14 per cent of workers require thoughtful retraining programs and support systems. Meanwhile, AI's growing energy footprint presents ongoing opportunities for innovation in sustainable computing and renewable energy integration. Artificial intelligence and ethical considerations The responsible development of AI requires addressing important ethical considerations with the same rigor we apply to technology itself. Workforce transition support becomes crucial as industries evolve, creating opportunities for new skilled roles while ensuring affected workers receive retraining and meaningful career paths. Environmental stewardship is equally important. While major tech companies have seen significant emissions increase largely due to AI infrastructure expansion, this challenge drives innovation in green computing and renewable energy solutions. Additionally, securing personal data becomes paramount as AI systems become more sophisticated, with recent studies showing 77 per cent of businesses experiencing AI-related data challenges, highlighting the need for robust privacy frameworks that protect individuals while enabling beneficial AI applications. Current data protection frameworks like Europe's GDPR provide excellent foundations, but AI's unique characteristics require evolved approaches. Unlike traditional data processing, AI systems integrate information in ways that make conventional privacy protections challenging to implement. This creates opportunities to develop next-generation privacy frameworks that address AI's permanent data integration while maintaining cross-border collaboration. The goal is to create comprehensive global standards that protect individual privacy while enabling AI's beneficial applications across borders. Understanding how personal data interacts with AI systems empowers users to make informed choices about their digital interactions. When individuals interact with ChatGPT or similar systems, they contribute to model training, creating opportunities for both personalized assistance and privacy considerations. This presents an opportunity to develop AI systems that provide personalized benefits while maintaining strong privacy protections through advanced techniques like differential privacy and federated learning. AI's extraordinary potential However, the story doesn't end with challenges - it begins with extraordinary potential. AI's capacity for human advancement continues to inspire remarkable breakthroughs. NASA's Perseverance rover uses AI for autonomous Mars exploration, opening new frontiers for space discovery. AI-powered stroke detection systems prove twice as accurate as traditional methods, saving lives through early intervention. The AI healthcare market has grown to over $32 billion, with applications reducing emergency room wait times and accelerating drug discovery for previously incurable diseases. These beneficial applications - from climate monitoring to space exploration - demonstrate AI's extraordinary capacity to address humanity's greatest challenges. The 57 countries that signed the Paris Declaration on AI governance last year recognize this tremendous opportunity. Building on this foundation, we can establish a global governing body that fosters international cooperation while ensuring AI development serves human flourishing. Like international frameworks for nuclear technology that enable both peaceful energy production and prevent harmful applications, AI governance can maximize benefits while addressing potential risks through shared standards and collaborative oversight. No doubt, the world needs responsible AI that can enhance our quality of life in all spheres and spaces. And the opportunity before us is clear: proactive governance now can unlock AI's full potential for a better tomorrow. That's the bottom line. (Krishna Kumar is a technology explorer & strategist based in Austin, Texas in the US. Rakshitha Reddy is AI developer based in Atlanta, US)
