GPT-6 Goes Wide: ChatGPT's Intelligent UI Reaches 1.2 Billion Weekly Users, and Free Plans Get It Today
OpenAI's Intelligent UI turns ChatGPT replies into charts, forms and tappable tools for its 1.2 billion weekly users, begins answering web-search queries 44% sooner than GPT-5.6 Instant, and opens to Free and Go plans on 8 October — the same day Anthropic ships a small model that costs 75% less to run.
More than 1.2 billion people open ChatGPT every week, and from today a large share of them get an answer that is no longer a wall of text. OpenAI's rollout of GPT-6 with Intelligent UI turns replies into charts, buttons, forms and small working tools you can click inside the conversation — and it is arriving on the free tier, not just the paid one.
What actually happened
OpenAI took GPT-6 from paid tiers to everyone. The company introduced the first GPT-6 models to paying customers last month; the announcement of 7 October extends the generation to a model built specifically for its largest audience, with Intelligent UI as the headline feature. Paid plans — Plus, Pro, Business and Enterprise — started receiving it in the Chat tab on 7 October, and Free and Go plans follow on 8 October, with enterprise rollouts governed by each company's admin settings. OpenAI confirmed the same rollout on its official X account.
Intelligent UI is a change of output format, not just tone. Rather than composing every response as prose, GPT-6 is trained to pick a layout per question: a side-by-side comparison where a comparison helps, an interactive diagram for an explanation, and plain text when plain text is the fastest route. Under the hood OpenAI describes a library of native, streamable components plus a compiler that processes the interface as the model writes it, so the interface appears progressively instead of after the whole answer is finished. Practical examples on the announcement page include a shopping-quantity calculator with a guest-count stepper and a dinner-bill splitter that divides items among named people.
The speed claim is specific, and vendor-reported. For questions that need a web search, OpenAI says GPT-6 Instant starts answering an average of 44% sooner than GPT-5.6 Instant, and that in an internal evaluation of everyday agentic tasks GPT-6 Extra High begins answering about as fast as GPT-5.6 Medium while scoring better than GPT-5.6 Extra High. The company also published a system card covering the safety evaluation for this release.
Which model you get depends on your plan. Paid tiers are served GPT-6 Sol, Free and Go tiers are served GPT-6 Luna, and the models behind ChatGPT Work and Codex are unchanged by this release. For builders, OpenAI's own model guide lists the family's API pricing: GPT-6 Astra at $10 input and $50 output per million tokens, GPT-6.1 Sol at $2 and $10, and GPT-6 Luna at $0.10 input and $0.50 output.
Secondary item of the day: Anthropic cut the price of its small model. Claude Haiku 5.5 arrived on 7 October as the company's cheapest, fastest small model, costing around 75% less to run than Haiku 4.5 — $0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100,000 tokens, against $1.00 and $5.00 on the previous generation. Anthropic reports 39.2% on Terminal-Bench 4.0 where Haiku 4.5 scored 0.0%, and availability across AWS, Google Cloud and Azure under the model id claude-haiku-5-5. Anthropic also halved Sonnet 5.5's cache-read price, which it says makes that model around 20% cheaper on most agentic work.
Why this matters for Indian SMBs and mid-market teams
- The free tier is now a legitimate working surface. Field staff, franchisees and consultants who will never be on a per-seat premium plan can use the same interactive answer format as the paid tier from today. For a distributor checking stock norms or a CA's office scaling a calculation across clients, "open ChatGPT and click" replaces "ask IT to build a small form".
- First-token latency is the metric that decides chatbot abandonment. A 44% earlier start on search-backed queries is the difference between a customer-facing assistant that feels instant on an Indian mobile connection and one that shows a spinner. Note the claim is OpenAI's internal evaluation — measure it against your own prompts before you quote it to a client.
- The cost floor keeps falling for high-volume work. GPT-6 Luna at $0.10 per million input tokens and Haiku 5.5 at the same $0.10 tier mean invoice-field extraction, ticket classification and message routing are now priced like infrastructure, not like consulting. At a few thousand tickets a day, model cost stops being the line item you argue about.
- Interactive output changes what you can skip building. A bill splitter, a parameter calculator, a comparison table with toggles — Intelligent UI generates these inside the chat. If your team's "internal tool" is a spreadsheet with a formula row, that tool may no longer need a developer; what it still needs is a trusted source of data and someone who owns the numbers.
- Enterprise admins keep the switch. Enterprise availability depends on workplace admin settings, so companies handling client data should decide deliberately whether ChatGPT is enabled, on which plan, and with what data controls — the model arriving on your employees' screens is not an IT procurement decision you get to make later.
Where Guruji Corporation fits in
Every one of these claims is vendor-published, and none of them is about your workload. We build the layer between a model launch and a working system: an evaluation harness that runs your real invoices, tickets and contracts against GPT-6 Sol, GPT-6 Luna and Haiku 5.5 so the "which model" question returns a cost and an accuracy number instead of an opinion; retrieval that grounds interactive answers in your own data rather than the model's memory; and the governance checklist — plan, data controls, logging — that turns a free-tier rollout into something your compliance team can live with.
If your team is already using ChatGPT informally, this week is the moment to make it official: pick the tasks where an interactive answer saves a human an hour, wire them to verified data, and price the whole thing at today's token rates. Talk to us and we will benchmark your actual use cases across GPT-6 and Haiku 5.5, show you what an Intelligent-UI-style output would replace in your current workflow, and give you a rollout plan your admin settings and your auditors both accept.
Sources
Want to see how we can build your project?
Send us your project outline or chat directly on WhatsApp.