What We Know
As of March 2026, OpenAI has not officially announced GPT-5 or its timeline. However, a combination of job postings, infrastructure investments, partnership disclosures, and carefully worded executive statements have assembled into a picture that the AI community has been piecing together. This article documents what is credibly sourced versus what is speculation.
Credible Signals
Several things are reasonably well-established: GPT-5 is in training as of Q1 2026 (inferred from Sam Altman's comment that "our next model will be significantly more capable" at Davos in January). The model is reportedly a native multimodal architecture — not text-first with multimodal bolted on. Compute investment for GPT-5 training is reportedly in the range of $2-4 billion, dwarfing any prior model's training budget.
Credible Leaks
Several individuals with verifiable access to OpenAI information have posted (and subsequently deleted) details suggesting: a 10 million token context window, native tool use and agent capabilities that don't require function calling workarounds, and real-time web access as a default mode rather than a plugin. These specifics should be treated with caution — they may reflect early capabilities that changed during training, or deliberate misdirection.
Timeline Speculation
The most commonly cited estimate from well-sourced AI journalists placed GPT-5 launch between April and July 2026. OpenAI's historical pattern of announcing major models 2-4 weeks before launch makes precise timeline prediction difficult. What is clear is that the AI race is intensifying: Anthropic's Claude 4 family was already available, Google's Gemini 2.5 had just launched, and the pressure on OpenAI to maintain its perceived leadership position with a flagship model update was growing.
What It Means for the Market
If GPT-5 delivers on even a fraction of the leaked capabilities, it will reset benchmark comparisons for the entire industry and likely trigger competitive responses from Anthropic and Google within 3-6 months. The most reliable planning assumption: by end of 2026, frontier AI capabilities will be substantially higher than today across all providers.
What This Means for Indian Businesses
GPT-5 is relevant for Indian businesses planning their AI strategy through 2026. If the leaked capabilities hold — particularly native agent architecture and expanded context — it could shift the economics of AI-powered professional services dramatically. Indian AI product companies should build on modular architectures that can switch underlying models without rebuilding, ensuring they can adopt GPT-5 (or any other breakthrough model) quickly when it arrives.