OpenAI upgrades ChatGPT with free reasoning for all users

The cost barrier that stopped many UK teams from exploring advanced AI reasoning has just been removed. On 6 August 2026, OpenAI upgraded ChatGPT across both its free and paid tiers, introducing more reliable outputs and new controls over how deeply the model thinks before it responds. For business owners weighing up AI costs and IT managers deciding which tools to roll out, this update changes the calculation.

What OpenAI actually changed

OpenAI made two distinct changes in this update. The first affects Plus and Pro subscribers, who now have access to an improved GPT-5.6 Sol model. This version produces more concise, factually reliable answers. It also introduces a customisable thinking slider, which lets users decide how much reasoning effort the model applies before generating a response.

The second change affects everyone on the free tier. Free users now get unlimited text chats and a new Think button. That button gives access to higher-reasoning capabilities for complex tasks, without requiring any paid subscription.

Before this update, deeper reasoning was locked behind a paid plan. Now any member of your team with a free ChatGPT account can access it.

Why this matters for UK SMEs

For most small and medium-sized businesses, AI adoption has followed a predictable pattern. A handful of staff on paid plans experiment with the tool. Everyone else waits to see whether the business will fund a wider rollout. This update disrupts that pattern.

Business owners no longer need to weigh subscription costs against headcount when deciding who gets access to capable AI. The free tier now does meaningful work. Staff across a 50-person team can use the Think button for drafting proposals, reviewing documents, or working through a complex planning task, at no additional cost.

For IT managers, the thinking slider on GPT-5.6 Sol addresses a different problem. It gives more control over how the model behaves in different contexts. A quick internal summary needs speed. A technical specification or a compliance document needs depth. The slider lets the user choose, rather than accepting a fixed default.

The improvement in factual reliability is also significant. OpenAI's own update notes that GPT-5.6 Sol delivers more reliable answers compared to previous versions. For UK SMEs where staff are using AI to draft client-facing content or analyse data, reducing factual errors matters. It shifts AI from a tool that needs heavy checking to one that earns a degree of trust.

AI adoption is 80% people and culture, and 20% technology. Removing the cost barrier does not automatically mean teams will use the tool well. That still requires guidance, clear use cases, and a way to build confidence.

Practical steps to take this week

1. Create a business account for ChatGPT - to protect your IP, comply to GDPR and monitor adoption

2. Ask those users to experiment with the thinking slider on a task that has previously produced unreliable AI output. Note whether the improvement is measurable and share the finding with the wider team.

3. Map two or three repeating tasks across your business where factual accuracy matters most, such as client proposals, supplier comparisons, or policy documents. These are the strongest candidates for an expanded free-tier rollout. Assign an IT manager or team lead to run a short test across those tasks before the end of the month.

Considerations and limitations

The free tier now does more than it did, but it still has limits. Free users access the Luna model, not the full GPT-5.6 Sol available to subscribers. For complex, high-volume tasks or integrations with other business tools, a paid plan will still deliver more consistent results.

The thinking slider gives useful control, but it also requires users to make a judgement call about when to use it. Without some guidance, staff may default to the fastest setting for every task, missing the benefit for work that genuinely needs deeper reasoning.

ChatGPT stores conversation data by default. IT managers should review OpenAI's data settings and confirm whether team members need to enable privacy controls before using the tool for sensitive business information.

Where to go from here

If you are a business owner or IT manager thinking about how to turn this update into a consistent habit across your team, that is exactly the kind of question the AI Readiness survey is designed to help you answer. Taking the survey gives you access to 65 or more free resources and a custom AI Readiness report. That report is the basis for a free 45-minute AI Readiness call to walk through the results and identify where to focus first.

gecco's Training and consultancy programme helps UK SMEs build the people-side of AI adoption, giving teams the practical skills and confidence to use tools like ChatGPT well, not just access them.

Website · LinkedIn · Case Studies · Newsletter