OpenAI Revamps GPT-5, Unveils ChatGPT Go In India & Altman Talks Future Plans

AI models have gone from academic tools to everyday conversational assistants. However, many users feel that they sound a little stiff and inorganic, with developers optimizing models to make them feel more amicable – without being sycophantic.  

Well, OpenAI’s newest release, GPT-5, was no exception with a launch that CEO Sam Altman admitted was “a little more bumpy than we’d hoped for.”
 
Some users said they preferred GPT-4o, describing GPT-5 as too formal or blunt.
In response, OpenAI is now tweaking that. In an X post, the company said it’s making GPT-5 “warmer and friendlier” after feedback that the system felt too formal before.


The changes are subtle, OpenAI explained, with small touches like “Good question” or “Great start” added to responses to make ChatGPT feel more approachable, with internal tests showing no increase in sycophancy compared to the GPT-5's earlier personality.

The update is expected to reach users within a day, with more refinements on the way. As VP Nick Turley put it, GPT-5 was previously “just very to the point,” but the goal now is to soften its tone and make interactions feel more natural.

Well, GPT has already shown it can handle everything from casual conversations to passing medical exams, but its next leap could reshape clinical decision-making. OpenAI has published a new paper positioning GPT-5 as a “generalist multimodal reasoner” for medicine, capable of combining patient narratives, lab data, and medical images into diagnostic reasoning.

In controlled evaluations, GPT-5 consistently outperformed both its predecessor GPT-4o and even pre-licensed human experts across major benchmarks, including MedQA, MedXpertQA, MMLU medical subsets, USMLE self-assessments, and VQA-RAD.

On MedXpertQA’s multimodal tasks, the model improved reasoning scores by +29.26% and understanding by +26.18% over GPT-4o, surpassing human experts by margins of +24.23% and +29.40%.

OpenAI’s case study highlights GPT-5 correctly identifying a complex diagnosis by integrating CT scans, lab results, and symptoms into a coherent reasoning chain and then recommending the next clinical step.

While these results mark a major advance, the paper stresses its benchmark conditions and that it doesn’t capture the uncertainty and ethical challenges of real-world medical practice.

Though GPT-5 is showing promise in fields like medicine, OpenAI’s leadership is already looking beyond the model itself. At a recent dinner with reporters in San Francisco, CEO Sam Altman gave a glimpse of the company’s bigger ambitions over bread rolls, fish entrées, and plenty of speculation about life after GPT-5.

Altman teased the company’s future from a forthcoming AI device designed with Jony Ive, to consumer apps under incoming applications chief Fidji Simo.


Its future moves could include an AI-powered browser to rival Chrome or even a new social platform. “If Chrome is really going to sell, we should take a look at it,” Altman said, though he doubted such a deal would materialize.

The conversation also touched on OpenAI’s interest in backing Merge Labs, a brain-computer interface startup to compete with Neuralink. “We have not done that deal yet; I would like us to,” Altman said.

Nick Turley, VP of ChatGPT, said the company is working with mental health experts to create a rubric for evaluating GPT-5’s responses, ensuring the model pushes back against unhealthy behaviors.

Despite the criticism, GPT-5 hasn’t slowed business as traffic doubled within 48 hours of launch, and Altman said demand has left the company “out of GPUs.” The dinner made one thing clear: OpenAI is preparing to grow well beyond ChatGPT by positioning itself as something closer to an Alphabet-scale giant.

ChatGPT may have hundreds of millions of users worldwide, but affordability has remained a sticking point in fast-growing markets like India. OpenAI is now addressing that with a new low-cost subscription tier, exclusively for India.

The AI startup introduced ChatGPT Go, a plan priced at ₹399 ($4.60) per month, less than a quarter of the existing Plus plan, which costs ₹1,999. Along with the lower price, OpenAI has also enabled local currency billing and payments via UPI, India’s largest payment network, making it easier for Indian users to subscribe.


ChatGPT Go includes everything in the free plan and adds more powerful features such as expanded access to GPT-5, legacy models (like 4o), and advanced tools such as deep research, agent mode, and even Sora video creation.

It also includes image generation, file uploads, and Python-powered data analysis. With longer memory, you’ll get more personalized chats, plus the ability to organize work into projects, manage tasks, and build custom GPTs.

Nick Turley said ChatGPT Go significantly expands what users get compared to the free tier, with “10 times” more access to messages, image generation, and file uploads. The plan also introduces longer memory for more personalized responses.

“Making ChatGPT more affordable has been a key ask from users! We’re rolling out Go in India first and will learn from feedback before expanding to other countries,” Turley said.

Do you think ChatGPT Go will become the entry point for millions of new AI users in India?
 
Can subtle tweaks in outputs make GPT-5 feel more human in everyday conversations?
 
Let us know your thoughts in the comments section below!