How to Run GPT-6.1 Terra in Parallel with Codex
GPT-6.1 Terra shipped in Codex on October 26, 2026. How to pick it in the Codex CLI, run several Terra sessions at once in separate Git worktrees, and do the same in Paradev.
For macOS on Apple Silicon and Intel. Linux AppImage is experimental.
Paradev was called Superset until October 2026.
To run GPT-6.1 Terra in parallel, start one Codex session per task, each in its own Git worktree and branch, and launch each with codex --model gpt-6.1-terra. In Paradev, create one workspace per task, choose Codex as the agent and GPT-6.1 Terra in the model picker, or pass --agent codex --model gpt-6.1-terra to superset ws create. Review each diff before you merge. Worktrees separate working files; they do not sandbox processes or prevent merge conflicts.
What shipped
On October 26, 2026, OpenAI released GPT-6.1 Terra. Its system card addendum calls it "a fast, efficient model for high-volume professional work, with strong image analysis and computer use capabilities." The model replaces GPT-5.6 Terra.
| GPT-6.1 Terra | |
|---|---|
| Released | October 26, 2026 |
| Where | ChatGPT Work and Codex (Plus, Pro, Enterprise, Edu), and the OpenAI API |
| API model id | gpt-6.1-terra |
| API price | $1 per million input tokens, $5 per million output tokens; OpenAI says standard API prices are 50% lower than GPT-5.6 Terra |
| Coding result OpenAI reports | 50.3% of tasks on a coding benchmark at maximum reasoning effort, against 21.5% for GPT-5.6 Terra |
| Codex CLI | In the model catalog from 0.165.0 (October 27); 0.166.0 (October 29) adds it to the Vertex AI and Amazon Bedrock catalogs |
OpenAI also says your Codex usage allowance goes further with GPT-6.1 Terra than with GPT-5.6 Terra. The benchmark and price figures come from OpenAI and are not independently checked.
Terra is not the Codex default. The Codex changelog lists no default model change, so a new session still starts on GPT-6.1 Sol unless you pick another model.
Why Terra fits parallel work
A cheaper, faster model changes how many sessions you can keep running. At lower prices per token, the same plan allowance or API budget covers more sessions. That makes Terra a sensible choice for the bounded tasks you fan out: tests, small fixes, docs, migrations with a clear spec. Keep the hardest task on the model you trust most and compare results before you switch everything.
Pick Terra in the Codex CLI
Update Codex to 0.165.0 or later, then choose the model when you start a session:
Inside a running session, /model switches models. Since 0.165.0, Codex also lets you change reasoning effort during a conversation on supported models.
Run several Terra sessions at once
Give each task its own worktree and branch, so two sessions never edit the same working files:
Open a terminal in each directory, install dependencies, and start Codex there:
Codex has used worktrees by default since 0.156, and 0.165.0 added a repository filter to the session picker and directory grouping in the agent command center, so sessions across linked worktrees are easier to find. Use narrow prompts with a clear stop condition:
Check each session with git branch --show-current before it starts editing.
What worktrees do not separate
Worktrees share repository history and run as your user. They do not give each task its own ports, database, credentials, or API quota. Two sessions can still change the same external service, and their branches can still conflict when you merge. Rerun your checks on the combined branch.
Run GPT-6.1 Terra in Paradev
Paradev is a desktop app that runs Codex and other CLI agents side by side. Each task runs in its own Git worktree and branch, with a terminal, a Changes pane, and a pull request pane next to it. Paradev does not proxy model calls: Codex runs on your own ChatGPT plan or API key. Paradev is open source under the MIT License.
From the app:
- Create a new workspace for each task.
- In the composer, choose Codex as the agent and GPT-6.1 Terra in the model picker.
- Write the task prompt and start it. Repeat for the next task.
- Review each diff in the Changes pane, then commit, push, and open a pull request from there.
From the superset CLI, start a batch of Terra workspaces and group them with a tag:
Tagged workspaces collect in a sidebar folder with the tag's name. --effort sets reasoning effort for the launch; omit it to use the Codex default. If you leave out --model, Codex starts on its own default model.
To run the same task on Terra and Sol and compare the diffs, create two workspaces with the same prompt and different --model values. The iPhone and iPad app lets you check on the sessions and review diffs while they keep running on your machine, a host you added, your own Linux server (Bring Your Own Cloud), or a cloud workspace.
Frequently Asked Questions
Is GPT-6.1 Terra the default model in Codex?
No. OpenAI's Codex changelog lists no default change with Terra. Select it with codex --model gpt-6.1-terra or /model, or pick it in Paradev's model picker.
Which Codex version do I need for GPT-6.1 Terra?
Codex CLI 0.165.0, released October 27, 2026, added Terra to the model catalog. Version 0.166.0 added it to the Vertex AI and Amazon Bedrock catalogs too.
Can I use GPT-6.1 Terra with an API key instead of a ChatGPT plan?
Yes. The API model id is gpt-6.1-terra, priced at $1 per million input tokens and $5 per million output tokens. In ChatGPT plans it is available to Plus, Pro, Enterprise, and Edu users.
Do parallel Terra sessions share files?
Not if each runs in its own Git worktree. They still share ports, databases, credentials, and your model quota, and their branches can conflict when merged.
Can I mix Terra with Claude Code or other agents?
Yes. In Paradev, each workspace runs the agent and model you choose, so a Codex session on Terra can run next to Claude Code, OpenCode, or Gemini CLI on the same repository.
Where can I read more about running Codex in parallel?
See the best IDE for Codex, Paradev vs Codex, and Git worktrees for AI agents.