Back to rankings

GPT-6.1 Terra Joins the Base44 Model Picker

Base44View original changelog

Base44 added OpenAI's GPT-6.1 Terra to its model picker, under More models, on both web and mobile. Users can set the model's effort level for each turn, as with other models that offer the setting. The model gives people who choose their own model another option for the AI chat.

Key Takeaways

  • GPT-6.1 Terra is now selectable in the Base44 AI chat model picker under More models.
  • The model is available on both web and the mobile app, so the choice follows users across devices.
  • Users can set a per-turn effort level, as with other supported models.
  • OpenAI released the model on October 26, 2026, so Base44 added it within days of launch.
  • OpenAI prices Terra at $1 input and $5 output per million tokens, half the price of its predecessor.
  • OpenAI still recommends GPT-6.1 Sol or GPT-6 Astra for the most demanding coding and research, so Terra suits everyday builds.

A New Model in the Picker

Base44 added GPT-6.1 Terra to the model picker in its AI chat. The model sits under the More models option, so it is available to people who choose their own model instead of leaving the choice to Base44. It appears on both the web app and the mobile app.

Effort Control Per Turn

As with other models that offer the setting, users can set the effort level of GPT-6.1 Terra for each turn. This lets a person trade speed and credit use against depth of reasoning on a message-by-message basis, for example using lower effort for a quick edit and higher effort for a harder debugging request.

Model Background

GPT-6.1 Terra is OpenAI's mid-tier model, released on October 26, 2026 as an upgrade to GPT-5.6 Terra. It is priced at $1 per million input tokens and $5 per million output tokens, which is 50 percent lower than its predecessor, and OpenAI describes it as faster at computer-use tasks. For the hardest coding and research work, OpenAI still points to GPT-6.1 Sol or GPT-6 Astra. Base44 has not published its own positioning for the model beyond adding it to the picker.

Why It Matters

For a vibe-coding platform, the model behind the agent is one of the most consequential variables. A cheaper, faster mid-tier option gives users a way to stretch credits on routine builds while keeping stronger models for complex work.