[News] GitHub Copilot Adds GPT-6.1 Sol, But Free Tier Users Are Locked Out

GitHub rolled out a new model inside Copilot on 2026-10-01: GPT-6.1 Sol, built by OpenAI and tuned specifically for agentic coding (where the AI runs multi-step tasks on its own, not just suggesting one line at a time) and terminal workflows. If you’re on the free tier wondering whether this changes anything for you, the short answer is no — this one sits behind a paywall higher than most people expect.

GitHub’s own changelog (dated 2026-09-29) says the model cuts token use and execution steps compared with GPT-6 and GPT-5.6. Tokens are the chunks of text a model processes to read your prompt and generate a reply — fewer tokens burned per task generally means cheaper, faster runs when a model is billed by usage. That matters because GPT-6.1 Sol is billed that way: usage-based, at whatever rate the model’s provider lists, not a flat “included” feature.

What GPT-6.1 Sol actually is

GitHub’s changelog frames this as an efficiency upgrade for agentic tasks — the kind where Copilot is asked to complete a multi-file change or run a sequence of terminal commands rather than just autocomplete a line. The model is now available in VS Code, Visual Studio, Copilot CLI, GitHub’s cloud coding agent, the GitHub mobile apps on iOS and Android, JetBrains IDEs, Xcode, and Eclipse. That’s a wide rollout footprint, but availability in your editor doesn’t mean availability on your plan — those are two separate gates.

Who gets it, who doesn’t

This is the part worth underlining: GPT-6.1 Sol is restricted to Copilot Pro+, Max, Business, and Enterprise subscribers. The plain US$10/month Pro plan does not get it, and neither does the free tier. If you’re paying for Pro today, this launch doesn’t add anything to your toolkit — you’d need to move up to Pro+ at US$39/month to touch it.

For teams, Business (US$19 per seat per month — a seat is one user license) and Enterprise (US$39 per seat per month) also get access, gated further by admin controls. Business and Enterprise admins can manage who sees the model through model policies in Copilot’s settings. By default the model turns on automatically for eligible accounts unless an admin disables it specifically or turns off new models generally. Because GitHub is rolling this out gradually, some eligible users may not see it in their model picker yet even if their plan qualifies.

🔍 AMPM exclusive check

We checked this against GitHub’s pricing and changelog pages rather than just repeating the vendor’s framing, and found two things that aren’t obvious from a quick skim of the changelog post.

First, the changelog lists “Pro+, Max, Business, and Enterprise” as eligible, but it’s easy to assume regular Pro (US$10/month) is included too since it’s the main paid individual tier. It isn’t. Our plan-by-plan check of GitHub’s own plan comparison confirms Pro explicitly excludes “premium models” as a category — GPT-6.1 Sol falls into that bucket, so Pro users are paying US$10/month and still don’t get this one.

Second, Business and Enterprise customers who signed up expecting extra AI Credits (prepaid usage allowances for premium-model requests, deducted per request) should know that GitHub’s temporary credit boost — 3,000 credits per seat for Business and 7,000 for Enterprise, instead of the standard 1,900 and 3,900 — ran from 2026-06-01 and ended 2026-09-01. As of today, 2026-09-30/10-01, that promotional window has closed and seats are back to standard allocations. Anyone budgeting for GPT-6.1 Sol usage under Business or Enterprise should plan around the standard 1,900 or 3,900 credits per seat, not the promo numbers that circulated over the summer.

Nothing else here contradicts GitHub’s own changelog — the IDE list, the gradual-rollout language, and the admin-toggle behavior all match what GitHub published.

☀️ AMO, the budget-minded cat

I don’t touch this one, and neither does anyone on the free tier. Free Copilot still gives you 2,000 code completions and 50 chat messages a month, and since 2026-06-24 free and student accounts can’t even pick a model manually — it’s locked to “Auto,” which chooses for you. So GPT-6.1 Sol isn’t a free-tier decision; it’s not on the menu at all.

Here’s the part that actually matters for anyone thinking about upgrading just for this model: Pro at US$10/month doesn’t unlock it either. You’d have to skip past Pro entirely and go to Pro+ at US$39/month — a US$29/month jump — just to access a model you may use occasionally for agentic tasks. If your work is mostly everyday autocomplete and simple chat, that’s a big price jump for one model. Business users should also note the seat-based AI Credits promo (3,000/seat instead of 1,900) ended 2026-09-01, so teams renewing now are back to standard credit math. My take: unless agentic, multi-step coding tasks are a real part of your week, there’s no financial case for upgrading tiers just to reach GPT-6.1 Sol.

🌙 PIMI, the performance-minded cat

This is the one I care about. GitHub says GPT-6.1 Sol is “tuned for agentic coding and terminal workflows” and uses fewer tokens and fewer execution steps than GPT-6 and GPT-5.6 to finish the same complex task reliably. If that holds up in real use, it means tasks like multi-file refactors or running a chain of terminal commands should complete faster and with less wasted computation per run — which, under usage-based billing, also means each task should cost less in credits than the same job on an older model.

The reach is solid too: it’s live across VS Code, Visual Studio, the Copilot CLI, GitHub’s cloud coding agent, JetBrains, Xcode, Eclipse, and even the mobile apps, so it’s not a desktop-only experiment. The catch is the gradual rollout — GitHub is staging availability, so even Pro+, Max, Business, or Enterprise users might not see it in their model picker on day one. If you’re on an eligible plan and want to try it for agentic work, check your model selector over the next few days rather than assuming it’s missing entirely.

Verdict: who should use it, who should skip it

Skip this one if you’re on Free, Student, or Pro — you either can’t select models manually or the plan simply doesn’t include premium models like this. Upgrading from Pro to Pro+ purely to get GPT-6.1 Sol only makes sense if agentic, multi-step coding tasks are a regular part of your workflow, since that’s exactly what the model is tuned for. If you’re already on Pro+, Max, Business, or Enterprise, it’s worth checking your model picker now — there’s no extra charge to try it beyond your existing AI Credits usage, and the lower token/step count claim means it could reduce your credit burn on complex tasks if it performs as described. Business and Enterprise admins renewing seats should also recalculate credit budgets now that the summer promo allocation has expired.

Written: 2026-10-01 Price last verified: 2026-09-30 Tool last health-checked: 2026-10-01

🔗 Related on AMPM

• GitHub Copilot pricing & free tier
• aider vs github-copilot
• bolt vs github-copilot

📎 Sources

All prices and quotas in this article come from AMPM’s verified dataset for GitHub Copilot (last verified 2026-09-30); primary sources:

• https://github.com/features/copilot/plans
• https://github.com/features/copilot

🐾 Meet the cats: AMO and PIMI

AMPM is a Taiwan-based AI-tool pricing watchdog. Our motto: ask before you subscribe. Two cats argue every tool from two sides:

• ☀️ AMO — the budget-minded cat. Always asks: is the free tier enough, and is paying actually worth it?
• 🌙 PIMI — the performance-minded cat. Cares about whether it works well, fast, and gets your job done.

When they are done arguing, you know whether to pay. Every number is checked by us, with the verification date at the end.

🐾 Our sister sites also checked

• Travel eSIM plans compared
• Exchange fees and safety checked