OpenAI adds o3 Pro to ChatGPT and drops o3 price by 80 per cent, but open-source AI is delayed
OpenAI described o3 Pro as a 'version of our most intelligent model, o3, designed to think longer and provide the most reliable responses.' The company notes that in internal and external academic evaluation, o3 Pro has consistently outperformed both o3 and its predecessors, especially in clarity, accuracy, instruction-following, and comprehensiveness. advertisement
Although o3 Pro uses the same underlying model as o3, it is optimised for dependability. OpenAI says reviewers have ranked the model higher across key tasks in science and data analysis. As OpenAI puts it, 'We recommend using it for challenging questions where reliability matters more than speed, and waiting a few minutes is worth the tradeoff.'
Notably, o3 Pro is also equipped with tools that enhance its capabilities, including the ability to browse the web, analyse documents, interpret visual inputs, run Python code, and use memory for more personalised interactions. However, these advanced tools mean its responses typically take longer to generate than those of o1 Pro.Despite the rollout, a few features remain unavailable in o3 Pro for now. Temporary chats are currently disabled due to a technical issue, and the model cannot yet generate images or use the Canvas interface. Users looking for image generation are advised to stick with GPT-4o, o3, or o4-mini.Enterprise and Edu customers will gain access to o3 Pro the following week.
A massive price cut for o3Alongside the launch of o3 Pro, OpenAI has also slashed the cost of using the o3 model by 80 per cent – a move that makes the model more accessible and affordable to developers. The cost of using o3 now stands at $2 per million input tokens and $8 per million output tokens, with additional discounts for cached (previously seen) prompts. This is a major reduction from the previous rates of $10 and $40 respectively. As OpenAI researcher Noam Brown pointed out on X, this update represents a significant shift in affordability for developers.OpenAI CEO Sam Altman confirmed the price drop on X, posting, 'we dropped the price of o3 by 80%!! excited to see what people will do with it now. think you'll also be happy with o3-pro pricing for the performance :)'
advertisementThe new pricing places OpenAI in sharper competition with rivals like Google DeepMind's Gemini 2.5 Pro, which charges between $1.25 and $2.50 for input and up to $15 for output tokens, depending on prompt size. Gemini also offers Google Search integration, but with extra fees beyond a certain usage limit.Anthropic's Claude Opus 4 remains the priciest competitor, charging $15 per million input tokens and $75 for output. DeepSeek's Reasoner and Chat models, on the other hand, lead the low-cost tier with rates as low as $0.035 for cached inputs during off-peak hours.Open-source model postponedWhile the launch of o3 Pro and the price drop for the o3 model are widely welcomed, OpenAI's open-source model has hit a delay. Originally expected in June 2025, the model will now arrive later, following a surprise development by the research team. The exact timeline for the release has not yet been disclosed. In a post on X, Altman explained, 'We are going to take a little more time with our open-weights model, i.e. expect it later this summer but not June.' He added that the delay is due to the team achieving something 'unexpected and quite amazing,' which still needs more time to polish.advertisementThe open-source model is expected to rival high-performing open reasoning models such as DeepSeek R1, with OpenAI aiming to raise the bar for open-access large language models.
OpenAI CEO reveals cost of a single ChatGPT queryAlongside these product updates, Altman has also addressed questions about the environmental cost of using ChatGPT. In a blog post published on Tuesday, he shared that each ChatGPT query consumes about 0.34 watt-hours, which is roughly what an oven uses in a second or a high-efficiency lightbulb in a couple of minutes.In terms of water usage, Altman revealed that the average query consumes 0.000085 gallons of water, or 'roughly one fifteenth of a teaspoon.' These figures, he explained, are part of a broader vision where 'the cost of intelligence should eventually converge to near the cost of electricity.'
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Mint
12 minutes ago
- Mint
Wall Street and AI Startups Are Fighting Over Entry-Level Quants
(Bloomberg) -- At a rooftop bar on Manhattan's Lower East Side, roughly 150 quant researchers met with employees at the artificial-intelligence startup Anthropic who implored them to consider a life away from Wall Street. Over plates of potstickers and popcorn chicken, they rubbed shoulders at the June mixer with former hedge fund quants-turned-Silicon Valley evangelists who encouraged them to apply for jobs at Anthropic, according to the company. This month, the San Francisco-based firm is going global with another quant 'social hour' in London. The recruiting campaign mirrors similar efforts by rivals OpenAI and Perplexity AI. And a number of leading figures in the AI industry already come with quant backgrounds, including OpenAI Chief Research Officer Mark Chen and Perplexity co-founder Johnny Ho. But quants lured by dreams of building AI models and tools instead of profit-seeking algorithms for traders — often for lofty pay and benefits comparable to the world of finance — also face the risk of disappointment. 'The pitch is 'come and build the machine god,'' said Agustin Lebron, a former Jane Street trader who now works at a systematic trading startup. 'But I suspect that, for a lot of those people, it'll end up being 'come and figure out how to make people buy things from ads.'' Still, the AI industry's competition with finance is noticeably heating up. For Wall Street, it's an unwelcome wrinkle in an already-brutal war for quant talent. Unlike financial firms, AI companies aren't covered by the non-competition agreements that keep many of these researchers from easily switching jobs. 'I'd estimate we've seen a 40–50% increase over the past 12–18 months in AI-native and software companies specifically asking for talent with quantitative finance backgrounds,' tech recruiter Mike Doonan said. Entry-level quants are eligible for base salaries of as much as $300,000, based on external job listings, but that doesn't include what can be considerable bonus targets. AI firms today can offer comparable base salaries, with compensation packages bolstered with equity rather than bonuses. According to an analysis of LinkedIn announcements, social media posts and company news sites conducted by employment tracking company Live Data Technologies, firms including Jane Street and Citadel Securities have lost quants to AI firms over the past year. Aron Thomas and Charles Guo both left Jane Street earlier this year to join Anthropic. In an interview, they praised their former firm as a great place to work but said they were drawn to the excitement of being part of the next big thing. 'It became very clear quite quickly that AI is going to change a lot of things and drive many changes in the world, and it seemed pretty important to be involved,' said Guo. Jane Street declined to comment for this story. Citadel also declined to comment on personnel matters, but pointed to growing interest in the firm's own internship program, which saw 108,000 applicants for this summer, up 20% from last year. While OpenAI also declined to comment, its chief executive officer Sam Altman touted quant-focused recruitment events in an April post on X. Noam Brown, an ex-quant and top researcher at the company, weighed in, noting that recruits don't need to take a pay cut anymore. Perplexity's Ho said the company pays $200,000 base salaries but makes up the difference somewhat with equity. He stressed that the firm's pitch to quants isn't mainly financial, though. Instead, it was the opportunity to take on 'new and more exciting challenges,' said Ho, who previously worked at Tower Research Capital. Quants are uniquely adept at minimizing latency in algorithms, which makes them desirable to AI developers competing to ensure users get responses as quickly as possible from the large language models that power generative AI tools such as ChatGPT. And much like AI research, quantitative trading involves sifting through vast amounts of unstructured data. What's more, firms like Anthropic and Perplexity are pushing more into financial services offerings. For its part, Anthropic said in a statement that it's after 'the rigorous analytical thinking and empirical research methods' that quants possess. Such skills have 'substantial overlap with the technical challenges of developing safer and more capable AI systems,' the company said, adding that it's going to keep hiring people with specialized backgrounds as it scales. Quant firms have occasionally sued employees departing for rivals. But litigation is unlikely for quants moving to AI labs, which don't directly compete with financial firms. Moreover, California — where most big AI companies are based — largely prohibits non-competition agreements. Ho said Wall Street has hurt itself with non-competes. 'They are becoming more and more secretive,' he said. There are signs that Wall Street is attempting to hit back against the poaching attempts of AI companies. For example, Iain Dunning, who oversees AI at Hudson River Trading, posted on X in May: 'Are you a researcher at OAI/Anthropic/etc and tired of overhiring, the orgchart chaos, the lowered talent bar, want to move to NYC, or just want to do something different?' 'Email me, DM me, mail a postcard,' he wrote. More stories like this are available on
Indian Express
12 minutes ago
- Indian Express
‘They use it to find an excuse': Israel is using Microsoft's Cloud to spy on Palestinians — and bomb them
Anything a Palestinian says can be used against them, quite literally. A phone call to a relative, a text to a friend, a message in the family group chat — any of those could be intercepted, analysed, and marked as suspicious by an Israeli military surveillance programme, built in part with Microsoft's cloud technology, and used as justification to bomb an area. A recent investigation by The Guardian, +972 Magazine, and Local Call reveals how Israel's elite cyber intelligence unit, Unit 8200, uses Microsoft's Azure cloud platform to store and process intercepted phone calls and texts from Palestinians in Gaza and the West Bank. Those calls are used not just for surveillance but for planning airstrikes. Unit 8200 has developed an AI tool, reportedly similar to ChatGPT, trained on these massive data troves. It can answer questions about individuals, offer insights from thousands of conversations, and flag potential targets. According to sources familiar with the system, this AI is being used to shape operations, justify detentions, and even assist in selecting bombing targets — sometimes based on who was near someone of interest. Israel has always intercepted Palestinian communications. But under this system, Unit 8200 stores all Palestinian phone calls for at least a month, allowing them to retroactively examine anyone's past conversations. The system began taking shape in 2021, when Unit 8200's commander, Yossi Sariel, met with Microsoft CEO Satya Nadella at the company's campus in Redmond. Sariel, who later resigned after the October 7 intelligence failure, pitched a plan to move Israeli military intelligence into the Azure cloud. Nadella reportedly backed the idea, telling Sariel: 'Microsoft is committed to providing resources to support.' Internal records, seen by The Guardian, show Nadella encouraged Unit 8200 to move up to 70 per cent of its data, including sensitive intelligence, into Microsoft's cloud. Microsoft now claims Nadella only attended the meeting for ten minutes and was unaware of the data's nature. A spokesperson said: 'We are not aware of Azure being used for the storage of such data.' But The Guardian obtained leaked Microsoft documents and spoke to eleven sources in the company and Israeli intelligence. They confirm that Azure has hosted massive quantities of Palestinian surveillance data — up to 11,500 terabytes, equivalent to 200 million hours of audio, some of it stored in Microsoft data centres in the Netherlands and Ireland. Sariel's vision, following a wave of lone-wolf Palestinian attacks in 2015, was to stop selective surveillance and instead monitor everyone. One system, still in use today, scans Palestinian text messages and flags ones containing 'suspicious' words like 'weapon' or 'death.' Another, nicknamed 'noisy message,' assigns a threat rating to every message sent in the West Bank. 'When they need to arrest someone and there isn't a good enough reason to do so, that's where they find the excuse,' said one Unit 8200 source to the Guardian. In Gaza, sources say the Azure-based system has been directly used to prepare airstrikes. Before hitting a target in a crowded area, officers examine recent calls made in the vicinity. Microsoft has said that its technology must not be used to identify bombing targets. But sources in Unit 8200 say that's exactly what is happening. Microsoft's ties to the Israeli military have sparked employee protests. At a conference in May, one employee interrupted Nadella's keynote yelling, 'how about you show how Israeli war crimes are powered by Azure?' After The Guardian exposed the company's role in January, Microsoft launched a review. The company has maintained it found 'no evidence to date' that its systems were used to harm civilians. But intelligence officers familiar with the system insist otherwise. They say it's helped kill thousands, including civilians and children. As of now, Microsoft continues to provide Unit 8200 with cloud infrastructure. Azure remains a 'mission-critical' part of Israel's surveillance machine, powering an AI-driven programme that listens to Palestinians, stores their voices, analyses their words, and sometimes helps choose who lives or dies.
Time of India
12 minutes ago
- Time of India
ChatGPT 5 available: Is it really free? How to access it and the 5 different price plans
A standout trait of OpenAI's GPT-5 is its refined emotional intelligence. The model can now interpret and respond to user input with a more human-like grasp of context and tone, enhancing its utility across diverse fields such as education, software development, mental health support, and entertainment. Tired of too many ads? Remove Ads ChatGPT 5: Revolutionising the AI Landscape What Makes ChatGPT 5 Different? Tired of too many ads? Remove Ads Multimodal capabilities: Users can share images, documents, and audio to receive relevant and timely responses. Users can share images, documents, and audio to receive relevant and timely responses. Enhanced memory: If enabled, GPT-5 can remember user preferences and past interactions, thereby personalising conversations. If enabled, GPT-5 can remember user preferences and past interactions, thereby personalising conversations. Smarter reasoning: Improvements in logical and analytical reasoning mean GPT-5 can now solve complex queries more accurately. Improvements in logical and analytical reasoning mean GPT-5 can now solve complex queries more accurately. Real-time responsiveness: Built for speed, GPT-5 significantly reduces latency, offering near-instantaneous replies even during heavy usage. Using GPT-5: Access Made Simple Is ChatGPT 5 Really Free? 1. Free Plan 2. Plus Plan ($20/month) 3. Pro Plan ($200/month) Tired of too many ads? Remove Ads 4. Team Plan ($25–$30/user/month) 5. Enterprise Plan (Custom Pricing) Openai gpt5: The Voice of the Future chatgpt 5: Personalisation at the Core OpenAI GPT5 : Bridging Accessibility and Innovation FAQs What is ChatGPT 5? Is ChatGPT 5 free to use? In a major stride towards the future of artificial intelligence, OpenAI has officially rolled out its latest and most sophisticated large language model, ChatGPT 5. With this launch, the California-based AI research company is redefining how users interact with AI-driven tools in everyday life, professional work, and creative endeavors. Branded by many in the tech space as OpenAI's most ambitious upgrade to date, GPT-5 is equipped to deliver deeper conversations, multimodal engagement, and a fluid user experience across launch arrives amid mounting anticipation, following the widespread adoption of previous GPT iterations. Designed atop the cutting-edge GPT-4o (omni) architecture, this version not only promises superior performance but also introduces a suite of enhancements that make artificial intelligence more conversational, intuitive, and newest iteration of OpenAI's language model stands out for its ability to handle context with increased precision. Whether the task involves interpreting complex images, engaging in live voice dialogue, or remembering user preferences across sessions, ChatGPT 5 offers a marked leap forward.A defining feature of OpenAI GPT5 is its nuanced emotional intelligence. The AI can now respond to user inputs with more human-like understanding, making it an effective companion for a range of use cases, from education and coding to therapy and latest release packs significant upgrades that go well beyond textual interaction. Chatgpt5 introduces:In addition, the interface has undergone a sleek redesign, aligning with OpenAI's goal of making AI interactions more user-friendly and less intimidating for new eager to try gpt-5 can do so through the official ChatGPT mobile app or desktop web interface. Once signed in, users simply need to select the GPT-4o model to engage with the most current access remains free and grants users the ability to interact with GPT-5's text and voice functionalities. However, more sophisticated tools such as file analysis, custom agents, image generation, and code debugging are housed behind paywalls, available through OpenAI's premium subscription plans, as per Open AI This question has lingered since the moment of announcement—and the answer is a nuanced currently offers five distinct pricing tiers to accommodate varied user needs:The most accessible tier offers a limited but functional experience. Users can chat with chat gpt 5 , use voice mode, and perform web-based queries. Some restrictions apply to the frequency and complexity of usage, especially during peak individual users who require more muscle, the Plus plan allows greater interaction volume, access to Sora video previews, and an expanded memory framework. Advanced voice features are also included in this toward developers, creators, and small businesses, the Pro tier includes full-scale access to GPT-5 Pro, extended computational support, and early access to experimental features. It also includes high-volume file processing, custom GPTs, and enhanced agent for collaborative settings, this plan is ideal for startups, research groups, or classrooms. Features include usage analytics, team file sharing, and admin controls. Annual billing at $25 per user offers savings over the monthly at large-scale deployments, this offering includes API access, high-security environments, 24/7 support, and custom infrastructure integration. Interested organisations must contact OpenAI directly for a tailored mode in openai gpt5 is no longer a gimmick. OpenAI has poured significant resources into developing lifelike voice interaction capabilities. Users can now hold conversations that feel natural, with the AI responding fluidly in real time. This feature is set to challenge existing voice assistants like Siri, Alexa, and Google Assistant, many of which rely on far less advanced models.A key selling point of chatgpt5 lies in its memory system. Once activated by the user, this memory stores contextual data such as preferred writing style, frequently asked questions, or past discussions. This level of customisation ensures that the AI grows more useful over time, a feature particularly advantageous for long-term research projects, therapy sessions, or academic OpenAI has reaffirmed its commitment to privacy. Memory can be turned off at any time, and users have control over what data is stored or offering GPT-5 under a freemium model, OpenAI attempts to balance accessibility with revenue generation. While critics argue that the most valuable tools remain paywalled, many in the tech community acknowledge the generosity of the free tier, especially in comparison to similar offerings from rival strategy not only broadens the reach of GPT-5 across the globe but also fosters AI literacy among the public, a mission that OpenAI has consistently promoted since its 5 is OpenAI's latest large language model, built on the advanced GPT-4o (omni) architecture. It introduces multimodal capabilities (text, voice, image, and file interaction), enhanced memory, and improved real-time reasoning, offering a more human-like and integrated AI ChatGPT 5 is available for free, but with limited access
