OpenAI has announced a subtle but notable update to its latest flagship model, GPT-5, aimed at making it “warmer and friendlier” after days of mixed user reactions to its launch.
The announcement, made late Friday, comes in response to feedback from users who said GPT-5 felt “too blunt” or “detached” compared to earlier versions such as GPT-4o. CEO Sam Altman himself admitted the rollout was “a little more bumpy than we’d hoped for,” acknowledging that OpenAI’s most advanced release to date had not immediately resonated with everyone.
In a social media post, the company said the changes would not alter the underlying reasoning power of GPT-5 but would instead make its interactions “more approachable.” According to OpenAI, users can expect “small, genuine touches like ‘Good question’ or ‘Great start,’ not flattery,” while internal tests showed “no rise in sycophancy compared to the previous GPT-5 personality.”
Register for Tekedia Mini-MBA edition 19 (Feb 9 – May 2, 2026): big discounts for early bird.
Tekedia AI in Business Masterclass opens registrations.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab: From Technical Design to Deployment (next edition begins Jan 24 2026).
At a dinner with journalists earlier in the week, OpenAI executives tried to shift attention to the company’s long-term strategy — including AI’s potential to reshape work, education, and even address looming demographic crises. But the uneasy launch of GPT-5 lingered in the background.
Vice President Nick Turley admitted that GPT-5, in its initial release, “was just very to the point,” a quality that frustrated some users who had grown used to the lighter tone and conversational fluidity of GPT-4o.
“This update will bring back some of that warmth,” Turley explained, framing the changes as a recalibration rather than a redesign.
The episode highlights a recurring pattern in OpenAI’s product evolution: every major release has faced a wave of pushback over tone and usability. When GPT-4o debuted earlier this year, some users complained it was “too chatty” and prone to filler language, prompting the company to tighten its style. Now, with GPT-5, the pendulum seems to have swung in the opposite direction, with users asking for more friendliness.
This back-and-forth underscores the difficulty of building AI systems that are both powerful and personable. For OpenAI, the challenge lies not only in advancing reasoning capabilities but also in meeting shifting user expectations about how AI should “feel.” Personality, it turns out, is just as important as raw intelligence when it comes to daily adoption.
Despite the rocky start, OpenAI remains the most highly valued AI company. Its most recent funding round placed its valuation at about $500 billion, far ahead of rivals like Anthropic, Mistral, and Google DeepMind. Analysts believe that even with occasional missteps, OpenAI continues to dominate adoption across industries, from enterprise software to healthcare applications.
The update is expected to serve as a test of whether GPT-5 can strike the right balance between intelligence and approachability. If it succeeds, the company could put its latest stumble behind it, reinforcing its position at the center of global debates over how AI should evolve — not just in terms of what it can do, but in how it relates to people.