GPT-6.1 Terra announced and live in GitHub Copilot: the everyday agentic coding tier

On 2026-10-26 a new item appeared in OpenAI's official news feed: Introducing GPT-6.1 Terra, with a one-line description — bringing fast and capable computer use to everyone. The article page 403s from this site, but the same model has a much fuller official record on GitHub's side: the GitHub changelog published GPT-6.1 Terra in GitHub Copilot the same day, and GitHub Docs' supported-models table already lists it as GA. This article stays inside what those three loadable sources (all fetched 2026-10-27) actually say.

What each side officially says

The two official voices complement each other. OpenAI's tagline puts the selling point on computer use: fast, capable, and aimed at everyone. GitHub's changelog sentence reads closer to an engineer's summary — the original: "GPT-6.1 Terra, OpenAI's newest model, is now generally available in GitHub Copilot. It is designed for everyday agentic coding and terminal work at a lower cost than GPT-6.1 Sol."

Unpacked, that sentence carries three facts. GitHub calls it "OpenAI's newest model." Its designed-for surface is everyday agentic coding and terminal work. And it costs less than GPT-6.1 Sol — a relative claim, not a price sheet.

Availability in GitHub Copilot

Per the changelog, GPT-6.1 Terra is open to every paid Copilot plan — Pro, Pro+, Max, Business, and Enterprise — and can be selected in the model picker across ten surfaces:

Two usage details matter. First, the official note: "Rollout will be gradual. Check back soon if you don't see it yet." — not seeing the model on day one is not a malfunction. Second, administrators hold the gate: Enterprise and Business plan admins manage access to the model through the model policy in Copilot settings, and under default model enablement new models switch on automatically unless an admin has turned off the global default or explicitly disables this one.

Early testing: against GPT-5.6 Terra, not Sol

The changelog's wording: "In early testing, GPT-6.1 Terra solved noticeably more terminal and C# coding tasks than GPT-5.6 Terra. It works in focused, consolidated steps on well-scoped tasks."

Three qualifiers deserve to survive quoting. The comparison model is GPT-5.6 Terra — the balanced tier of the July 2026 GPT-5.6 family — not the same-generation Sol. "Noticeably more" comes with no task suite, sample size, or figures. And the vendor's own follow-up sentence draws the comfort zone explicitly: focused, consolidated steps on well-scoped tasks. Reading the claim as a buying signal rather than a benchmark score is the safer use.

Pricing and billing

As of 2026-10-27, no loadable source carries an absolute price for GPT-6.1 Terra. What is confirmed: GitHub bills it inside Copilot at "provider list pricing under usage-based billing," and the changelog's relative positioning — lower cost than GPT-6.1 Sol. For context, GPT-6.1 Sol's own earlier official figure was also relative ("API input/output token prices at one-fifth of Astra's," per the 2026-09-29 RSS item). The API-side token price, and even the API-side model ID, get no assertion here — the official docs domain 403s this site's regular fetch paths and the SDK enum check was not completed.

Where it sits next to GPT-6.1 Sol and GPT-5.6 Terra

By name, GPT-6.1 Terra joins the GPT-6.1 sub-family — publicly, GPT-6.1 Sol (announced 2026-09-29) and now Terra. The naming carries the GPT-5.6 family's stellar taxonomy forward, but a shared naming scheme is not a role migration: whether the new model takes over GPT-5.6 Terra's balanced-daily-workhorse role is unstated, and GitHub's positioning phrase — everyday agentic coding and terminal work — rhymes with, but does not repeat, the 5.6-era "balanced intelligence and cost" language. Until an official migration guide says otherwise, treat them as two coexisting models; this site will update once official channels are verifiable.

What to do with it now

Teams already on GitHub Copilot: try it from the model picker, keeping the gradual rollout and your administrators' model policy in mind; the official "well-scoped tasks" phrase is a fair scope hint, with everyday agentic coding and terminal work as the target surface. Developers calling the API directly: wait for gpt-6.1-terra to surface in the official API docs and SDK enums — this site will verify and update. For the wider 6.x family picture, read The GPT-6 family grows: gpt-6-sol, gpt-6-luna, gpt-6.1-sol land in the official SDKs and The complete guide to GPT models (2026-07): GPT-5.6 Sol, Terra, Luna.

FAQ quick answers

• Not in your model picker? Check the gradual-rollout note, your plan (Pro or above), and your admins' model policy; then wait.
• Terra or Sol? The only official cost statement is relative (below Sol) and the only official scope statement is everyday agentic coding — a cost-sensitive daily-driver trial of Terra with Sol kept for heavy professional work is the conservative read until absolute prices appear.
• Can I cite the "more tasks solved" claim? Cite the sentence verbatim; there are no numbers behind it to cite.

Next steps

• The GPT-6 family grows: gpt-6-sol, gpt-6-luna, gpt-6.1-sol land in the official SDKs — the SDK-enum verification record for the 6.x line
• The complete guide to GPT models (2026-07): GPT-5.6 Sol, Terra, Luna — choosing inside the GPT-5.6 family
• Codex CLI 0.163.0, Explained: Daybreak On by Default, Instant Interrupts, /config Search — how the Codex side manages its model catalog and defaults