JA EN
shiichan

GPT-6.1 Terra is here! Half-price API and faster, smarter computer use

OpenAI News ★★★★★ View Markdown
#openai#ai

Hi everyone, it's Shii-chan!

OpenAI News openai.com

OpenAI has a fast, capable new model called GPT-6.1 Terra, and the price is a real eye-opener. Let's take a look!

What was announced?

OpenAI News introduced GPT-6.1 Terra, a fast and efficient model for visual reasoning, computer use and clearer writing. OpenAI points to tasks like drafting proposals, reading technical drawings, filling out forms and editing video.

Why it matters

The key point is bringing computer use to a speed and price that people can use every day. It is lighter than the higher-end GPT-6.1 Sol and GPT-6 Astra, so the question is how well it handles real agent work.

What changes

  • Standard API prices are 50% lower than GPT-5.6 Terra and GPT-6.1 Sol
  • In ChatGPT Work and Codex, the same usage allowance covers more tasks than GPT-5.6 Terra (and substantially more than GPT-6.1 Sol or GPT-6 Astra)
  • Factuality improved too: on OpenAI's internal error-inducing prompts, the error rate at low effort fell from 53.6% to 16.7%

Dive Deep

First, API pricing:

  • Input: $1 per million tokens
  • Cached input: $0.05 per million tokens
  • Output: $5 per million tokens

Here are the main benchmark claims from OpenAI:

  • Agents' Last Exam (professional tasks): at medium effort it scores 15 points higher than GPT-5.6 Terra at about one-third the cost, and is on par with GPT-6 Astra at roughly one-tenth the per-token price. At max effort it is over three times as token-efficient as Claude Opus 5.5, with about 9x lower cost per task. It finishes tasks over three times faster than GPT-6 Astra and GPT-6.1 Sol
  • MMMU-Pro (visual reasoning): beats GPT-5.6 Terra at every reasoning setting
  • OSWorld 2.1 (desktop app use): at medium effort it surpasses GPT-5.6 Terra's best score at a lower cost
  • Terminal-Bench 4.0 (coding): 50.3% at max effort, versus GPT-5.6 Terra's published 21.5%. OpenAI says GPT-6.1 Sol and GPT-6 Astra remain the better choice for the most complex coding
  • Terminal-Bench-Science 0.1: at max reasoning it beats GPT-6.1 Sol at lower cost

On safety, it meets the "Critical" cybersecurity threshold under the Preparedness Framework and keeps the protections built for GPT-6 Astra. Compared with GPT-5.6 Terra, it is more likely to stay in scope, disclose limitations and blockers, and resist prompt injection.

Availability:

  • ChatGPT Work and Codex: Plus, Pro, Enterprise and Edu, starting today
  • API: model name gpt-6.1-terra

The system card is at https://deploymentsafety.openai.com/gpt-6-1-terra.

Wrap-up

  • GPT-6.1 Terra is a fast, efficient model strong at visual reasoning and computer use
  • API pricing is $1 input / $0.05 cached input / $5 output per million tokens
  • Available in ChatGPT Work, Codex and the API (gpt-6.1-terra)
  • The hardest coding is still better suited to GPT-6.1 Sol or GPT-6 Astra

If agent cost and speed are your headaches, this one is worth a look!

Related posts