OpenAI Unveils GPT-5, Eyes $500 Billion Valuation Amid AI Race

[justseyi]

OpenAI launched its highly anticipated GPT-5 model on August 7, 2025, during a livestreamed event led by CEO Sam Altman, marking a pivotal moment for the AI pioneer behind ChatGPT. Described as a “PhD-level expert” in performance, GPT-5 promises smarter, faster, and more human-like interactions, bringing the sci-fi vision of Her closer to reality. Amid the rollout, OpenAI is reportedly negotiating a secondary stock sale that could value the company at $500 billion, a 67% jump from its $300 billion valuation in March 2025, per Bloomberg. As competition intensifies with rivals like Google DeepMind and xAI, GPT-5’s debut underscores OpenAI’s bid to maintain its AI dominance while navigating valuation pressures and ethical concerns.

GPT-5: A Leap Toward Human-Like AI

Showcased via OpenAI’s YouTube channel, GPT-5 integrates the reasoning prowess of OpenAI’s o-series models with the speed of its GPT series, offering a unified model that auto-selects the best response mode—quick replies or in-depth reasoning—without user input. Altman, speaking at the launch, compared GPT-5 to “talking to a PhD-level expert,” a leap from GPT-4’s “college student” capabilities. Key demonstrations included:

• Vibe Coding: GPT-5 created a polished French-learning game from a natural language prompt in under a minute, raising concerns at Duolingo about AI-driven educational tools. On the SWE-bench Verified coding test, it scored 74.9%, edging out Anthropic’s Claude Opus 4.1 (74.5%) and Google DeepMind’s Gemini 2.5 Pro (59.6%).

• Human-Like Responses: The model’s written and voice outputs shed robotic tones, making ChatGPT-generated content harder to detect. It scored 89.4% on the GPQA Diamond test for PhD-level science questions, surpassing Claude Opus 4.1 (80.9%) and xAI’s Grok 4 Heavy (88.9%).

• Enhanced Functionality: With user permission, GPT-5 integrates with Gmail, Google Calendar, and Google Contacts to generate comprehensive daily schedules, acting as a proactive assistant.

• Reduced Hallucinations: GPT-5 hallucinates only 4.8% of the time on ChatGPT prompts, compared to 22% for o3 and 20.6% for GPT-4o, and just 1.6% on health-related queries, per HealthBench Hard Hallucinations.

Available to all ChatGPT users—free tier with capped access, Plus ($20/month) with higher limits, and Pro ($200/month) with unlimited GPT-5 and GPT-5 Pro access—GPT-5 also offers four new chat personalities (Cynic, Robot, Listener, Nerd) for tailored interactions. Developers can access GPT-5 via OpenAI’s API in three sizes—gpt-5, gpt-5-mini, and gpt-5-nano—priced at $1.25 per million input tokens and $10 per million output tokens.

Mixed Reception and Industry Impact

Early reviews, reported by Mashable, praise GPT-5’s speed and conversational finesse but note it’s “not that much better” than GPT-4, with incremental improvements rather than revolutionary leaps. On Tau-bench, testing online task navigation, GPT-5 scored 63.5% for airline websites (below o3’s 64.8%) and 81.1% for retail sites (below Claude Opus 4.1’s 82.4%). Still, its coding and health response accuracy, with “PhD-level” expertise, positions it as a versatile tool for developers, educators, and consumers.

Altman, in a Theo Von podcast, likened GPT-5’s development to the Manhattan Project, admitting it “scares” him due to its speed and capabilities, which once made him feel “useless” when it effortlessly handled a complex email. He told reporters GPT-5 is “the best model in the world,” a claim tempered by its mixed benchmark performance against rivals. NBC News called GPT-5 a “barometer” for AI progress, noting its launch tests whether generative AI is plateauing or advancing rapidly.

$500 Billion Valuation Push

The GPT-5 launch coincides with OpenAI’s negotiations for a secondary stock sale, allowing employees to cash out at a $500 billion valuation, up from $300 billion in a $40 billion SoftBank-led round in March 2025. Investors like Thrive Capital are eyeing billions in employee shares to capitalize on OpenAI’s growth, fueled by ChatGPT’s 700 million weekly users and 3 billion daily messages. The sale aims to retain talent amid poaching by Meta Platforms, offering $1.5 million bonuses over two years, per Benzinga.

OpenAI’s valuation surge, potentially surpassing SpaceX’s $400 billion, reflects investor confidence but faces hurdles. Negotiations to transition from a nonprofit to a public benefit corporation are stalled, with Microsoft, a $13.75 billion investor, seeking to secure its stake through 2030. Regulatory scrutiny from California and Delaware attorneys general, plus a lawsuit from co-founder Elon Musk, complicates the shift.

Challenges and Ethical Concerns

OpenAI warned of potential “hiccups and capacity crunches” as GPT-5’s computational demands strain infrastructure. Anthropic revoked OpenAI’s API access, alleging terms violations, while Getty Images’ chief product officer raised concerns about GPT-5’s training data authenticity, urging creator protections. OpenAI’s new “safe completions” approach reduces deception, with GPT-5 refusing unsafe queries while explaining limits, enhancing user trust, per VP Alex Beutel.

Looking Ahead

GPT-5’s rollout, accessible to all ChatGPT users and developers, sets a new standard for AI, with SiliconRepublic calling it “expert-level intelligence in everyone’s hands”. Its integration with Google services and coding prowess could disrupt industries, from education to software development, though modest benchmark gains temper expectations. As OpenAI navigates its $500 billion valuation and corporate restructuring, follow updates on CBS News and OpenAI’s X page to see how GPT-5 shapes the AI landscape in 2025.

[Author]

Posts