OpenAI has launched GPT-5, a new flagship AI model that will power the company\u2019s next generation of ChatGPT. <\/p>\n
GPT-5, which was released Thursday, is OpenAI\u2019s first \u201cunified\u201d AI model and combines the reasoning abilities of its o-series<\/a> of models with the fast responses of its GPT series<\/a>. The next-generation model signals a new era for ChatGPT \u2014 and its creator, OpenAI \u2014 pointing to OpenAI\u2019s broader ambitions to develop AI systems that are more like agents<\/a> than chatbots.<\/p>\n While GPT-4<\/a> enabled AI chatbots to offer smart responses on a wide variety of questions, GPT-5 allows ChatGPT to complete a wide variety of tasks on behalf of users \u2014 such as generating software applications, navigating a user\u2019s calendar, or creating research briefs. <\/p>\n With GPT-5, OpenAI has also sought to make ChatGPT simpler to use. Instead of asking users to choose the right settings, GPT-5 comes equipped with a real-time router that decides how to offer the best answer, whether that\u2019s responding to user questions quickly or taking additional time to \u201cthink\u201d through answers.<\/p>\n During a briefing with reporters, OpenAI CEO Sam Altman claimed GPT-5 is \u201cthe best model in the world,\u201d and said it represented a \u201csignificant step\u201d along the company\u2019s path to developing AI that can outperform humans at most economically valuable work \u2014 that is, artificial general intelligence (AGI).<\/p>\n \u201cHaving something like GPT-5 would be pretty much unimaginable at any previous time in history,\u201d said Altman.<\/p>\n Starting Thursday, GPT-5 will be available to all free users of ChatGPT as their default model. OpenAI\u2019s VP of ChatGPT, Nick Turley, said this is part of the company\u2019s effort to give free users access to an AI reasoning model for the first time. (Previously, the company gated these more advanced models behind a paywall.)<\/p>\n \u201cThis is just one of the ways that I\u2019m excited to live the mission, making sure that this stuff actually benefits people,\u201d said Turley on the decision, referencing OpenAI\u2019s long-standing mission<\/a> to distribute advanced AI to as many people as possible.<\/p>\n Techcrunch event<\/p>\n \n\t\t\t\t\t\t\t\t\tSan Francisco<\/span> The expectations are high for GPT-5, one of OpenAI\u2019s most anticipated product launches since ChatGPT put the company on the map in 2022. Since then, ChatGPT has grown into one of the world\u2019s most popular consumer products, reaching more than 700 million users every week \u2014 nearly 10% of the globe\u2019s population, according to the company.<\/p>\n Many see GPT-5 as a bellwether for AI progress broadly, and the model\u2019s reception by Silicon Valley could have profound implications for Big Tech, Wall Street, and policymakers regulating technology. These stakeholders are watching to see if GPT-5 offers a significant jump in AI\u2019s capabilities, much like its predecessor, GPT-4, which challenged expectations of what software can do.<\/p>\n OpenAI claims GPT-5 is state-of-the-art in several domains, slightly edging out leading AI models from Anthropic, Google DeepMind, and Elon Musk\u2019s xAI on key benchmarks. However, GPT-5 slightly underperforms frontier AI models in other areas.<\/p>\n The company says GPT-5 offers frontier-level performance around coding; Altman said the model specifically excels at spinning up entire software applications on demand, in what\u2019s become known as \u201cvibe coding.\u201d<\/p>\n On SWE-bench Verified \u2014 a test of real-world coding tasks pulled from GitHub \u2014 GPT-5 scores 74.9% on its first attempt. That means GPT-5 just outperforms Anthropic\u2019s latest Claude Opus 4.1 model, which scored 74.5%<\/a>, and Google DeepMind\u2019s Gemini 2.5 Pro, which scored 59.6%<\/a>.<\/p>\n On Humanity\u2019s Last Exam \u2014 a difficult test measuring AI model performance across math, humanities, and the natural sciences \u2014 a version of GPT-5 with extended reasoning (GPT-5 Pro) scored 42% when using tools. That\u2019s slightly less than xAI was able to achieve with Grok 4 Heavy<\/a>, which scored 44.4% on the test.<\/p>\n On GPQA Diamond \u2014 a test of PhD-level science questions \u2014 GPT-5 pro scored 89.4% on its first try, outperforming Claude Opus 4.1, which scored 80.9%, and Grok 4 Heavy, which scored 88.9%.<\/p>\n OpenAI says GPT-5 is better for answering health-related questions. On a test measuring accuracy in AI model responses around healthcare topics, HealthBench Hard Hallucinations, OpenAI says GPT-5 (with thinking) hallucinates just 1.6% of the time. This is far lower than the company\u2019s previous GPT-4o and o3 models, which scored 12.9% and 15.8, respectively. <\/p>\n While AI chatbots are not medical professionals, millions of people<\/a> are using them for health advice. In response to this phenomenon, the company says GPT-5 is more proactive about flagging potential health concerns and helping users parse medical results.<\/p>\n In addition, OpenAI says GPT-5 is better than other AI models on more difficult-to-measure, subjective domains,<\/a> such as creative design and writing. Turley said GPT-5 responds more naturally and exhibits \u201cbetter taste\u201d than other AI models on creative tasks.<\/p>\n \u201cThe vibes of this model are really good,\u201d said Turley.<\/p>\n GPT-5 is also more accurate than OpenAI\u2019s previous models, and the company says it suffers far less from hallucinations \u2014 the tendency for AI models to make up information \u2014 compared to its o-series models. Hallucinations seemed to be getting worse<\/a> in OpenAI\u2019s latest AI reasoning models, such as o3, and OpenAI previously said it didn\u2019t quite understand why it was happening.<\/p>\n In responses to ChatGPT prompts, OpenAI found that GPT-5 (with thinking) hallucinates and responds with incorrect information 4.8% of the time. That\u2019s a significant reduction from o3 and GPT-4o, which score hallucination rates of 22% and 20.6%, respectively, on the test.<\/p>\n On a benchmark measuring an AI model\u2019s agentic ability to complete simulated online tasks, Tau-bench, GPT-5 offers mixed performance. On part of the test measuring an AI\u2019s ability to navigate an airline\u2019s website, GPT-5 scores 63.5%, slightly underperforming o3, which scored 64.8%. On another part of the test measuring AI\u2019s ability to navigate retail websites, GPT-5 scores 81.1%, underperforming Claude Opus 4.1, which scored 82.4%.<\/p>\n OpenAI also says that GPT-5 is safer than its previous models. While AI reasoning models occasionally exhibit a tendency to scheme against humans<\/a> or lie to promote their own goals, OpenAI found that GPT-5 was deceptive at a lower rate than other models. <\/p>\n Alex Beutel, OpenAI\u2019s safety research lead, said reducing deception improves not only the safety of GPT-5, but also the user experience, creating a model that\u2019s more \u201ctransparent and honest in ways users can trust.\u201d<\/p>\n Beutel also notes GPT-5 is better at discerning between bad actors who are trying to misuse ChatGPT and users making harmless requests. This results in GPT-5 being able to refuse more unsafe questions, while offering fewer rejections to users seeking harmless information.<\/p>\n ChatGPT is getting a few user experience upgrades as part of the GPT-5 launch. Users can now select from four new personalities in ChatGPT\u2019s setting: Cynic, Robot, Listener, and Nerd. The company says these will adapt ChatGPT\u2019s responses without requiring users to specifically ask the model to respond in a certain way.<\/p>\n Subscribers to ChatGPT\u2019s $20-per-month Plus plan get higher usage limits for GPT-5 than free users. Meanwhile, $200-per-month Pro subscribers will have unlimited access to GPT-5, as well as a souped-up version called GPT-5 Pro that uses additional computational resources to produce better answers. Organizations on OpenAI\u2019s Team, Edu, and Enterprise plans will gain access to GPT-5 as their default model next week.<\/p>\n For developers, GPT-5 is coming to OpenAI\u2019s API in three sizes \u2014 gpt-5, gpt-5-mini, and gpt-5-nano \u2014 which will spend more or less time \u201creasoning\u201d through tasks. Developers can also now control verbosity in the OpenAI API, deciding how long or short an AI model\u2019s responses should be.<\/p>\n The base model of GPT-5 will cost developers $1.25 per million input tokens (roughly 750,000 words, longer than the entire \u201cLord of the Rings\u201d series) and $10 per million output tokens.<\/p>\n The launch of GPT-5 comes after a busy week for OpenAI. The company released an open-weight reasoning model, gpt-oss<\/a>, that developers and enterprises can download for free and run at a fraction of the cost. The open model nearly matched the abilities of OpenAI\u2019s previous top models, o3 and o4-mini, but GPT-5 sets a new standard for frontier performance in some areas, such as coding.<\/p>\n However, GPT-5 seems to be roughly on par with other frontier AI models in several areas. Benchmarks, of course, only tell part of the story for any AI model, and it remains to be seen how developers will use GPT-5 in the real world, and whether the model is truly a step above the competition.<\/p>\n<\/div>\n
\n\t\t\t\t\t\t\t\t\t\t\t\t\t|<\/span>
\n\t\t\t\t\t\t\t\t\t\t\t\t\tOctober 27-29, 2025<\/span>\n\t\t\t\t\t\t\t<\/p>\n<\/p><\/div>\n<\/p><\/div>\n<\/div>\nGPT-5 offers a slight edge on the competition<\/h2>\n
![\"\"](\"https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/08\/Screenshot-2025-08-07-at-12.56.46PM.png?w=680\")
Upgrades for consumers and developers<\/h2>\n