OpenAI has rolled out GPT-6.1 Terra, a new mid-range model built less for conversation and more for getting work done on a computer. Launched on 26 October 2026, Terra can read what is on a screen, click, type and move between applications to finish multi-step tasks, the same way a human operator would.

The release signals how quickly the industry’s focus is shifting from chatbots to agentic AI that can act inside existing enterprise software, a shift that matters for Indian IT services firms, global capability centres and enterprises running large back-office operations.

From answering questions to operating software

Unlike earlier text-first models, GPT-6.1 Terra is designed for “computer use”. It visually interprets an interface and works through it step by step. OpenAI positions it for routine but time-consuming jobs such as moving invoice data into accounting software, completing long online forms, drafting business proposals and even turning architectural drawings into 3D models.

“6.1 Terra is here! It is good at a lot of things, but we are especially excited for this new era of computer use,” OpenAI CEO Sam Altman said on X while announcing the model.

Faster, cheaper and fewer errors

OpenAI says Terra completes tasks roughly three times faster on average than its heavier models, GPT-6.1 Sol and GPT-6 Astra, and matches GPT-6 Astra on its internal Agents’ Last Exam benchmark at about one-tenth of the price per token. These are the company’s own figures and are yet to be independently verified.

Pricing is a key part of the pitch. API rates have been cut by half compared with GPT-5.6 Terra, to $1 per million input tokens and $5 per million output tokens, down from $2 and $12. For enterprises automating workflows at scale, that brings agent-style automation closer to everyday economics.

OpenAI also reports a sharp improvement in reliability. In its internal tests, the error rate on simple tasks fell from 53.6 per cent with the older model to 16.7 per cent, and to 4.3 per cent at maximum reasoning effort. The model more than doubled its predecessor’s score on complex terminal-based coding benchmarks and is already reaching developers through tools such as GitHub Copilot.

Where Terra fits

OpenAI’s current model family spans Luna, a lightweight option; Terra, the mid-range workhorse; and Sol and Astra for heavy-duty, complex work. With this release, Terra joins Sol and Luna in the GPT-6.1 generation.

GPT-6.1 Terra is available to ChatGPT Plus, Pro, Enterprise and Edu users through ChatGPT Work and Codex, via OpenAI’s API as gpt-6.1-terra, and on Amazon Bedrock.

Why it matters for India

For Indian enterprises and technology service providers, cheaper and more dependable computer-use agents could reshape process automation, testing and IT operations work that has traditionally relied on large teams. At the same time, the launch raises familiar questions for CIOs around governance, access controls and human oversight when AI agents are allowed to operate business applications directly.

Photo: Sam Altman, CEO, OpenAI. Credit: TechCrunch / Wikimedia Commons (CC BY 2.0).

Be a part of Elets Collaborative Initiatives. Join Us for Upcoming Events and explore business opportunities. Like us on Facebook , connect with us on LinkedIn and follow us on Twitter.

"Exciting news! Elets technomedia is now on WhatsApp Channels Subscribe today by clicking the link and stay updated with the latest insights!" Click here!