Free $0, Pro $10/mo, Pro+ $39/mo, Business $19/user. Since June 2026 Copilot bills usage as token-based AI Credits. Here's the full breakdown — and how to cut it.
Every plan includes code completions plus a monthly allotment of usage-based AI Credits. Here's what each tier costs and what's included in 2026.
| Plan | Price | Included AI Credits | Completions | Agent Mode / Chat | Best For |
|---|---|---|---|---|---|
| Free | $0 | Small monthly allotment | Included (no credits) | Limited — draws from credits | Trying Copilot, hobby use |
| Pro | $10 / mo | Monthly allotment included | Included (no credits) | Draws from credit pool | Most individual developers |
| Pro+ | $39 / mo | Larger monthly allotment | Included (no credits) | Frontier models + more credits | Heavy agent-mode / chat users |
| Max | $100 / mo | Largest individual allotment | Included (no credits) | Maximum credits + top models | Power users, all-day agents |
| Business | $19 / user/mo | Includes $19/mo in AI Credits | Included (no credits) | Team admin, policy controls | Teams & small orgs |
| Enterprise | $39 / user/mo | Includes $39/mo in AI Credits | Included (no credits) | SSO, audit, org models | Large orgs, compliance needs |
Code completions and Next Edit suggestions never consume AI Credits. Business & Enterprise plans get 2x promotional AI Credits through August 2026 during the transition.
This is the single biggest thing to understand about Copilot pricing in 2026. On June 1, 2026, GitHub moved from counting "premium requests" to usage-based AI Credits billed on real token consumption.
Every plan includes a monthly allotment of AI Credits, and paid plans can buy more. Code completions and Next Edit suggestions do not consume AI Credits — they stay part of your base subscription. What does draw from your credit pool is agent mode, chat, and code review. Usage is calculated on token consumption (input + output + cached tokens) at each model's listed API rates. Once your monthly credits are gone, further usage is billed as overage at API token rates. That's why heavy agent-mode users on frontier models can blow past their base subscription cost.
Your base plan is only the floor. Under AI Credits, what you actually pay depends on how hard you lean on agent mode and chat. Here's what that looks like across usage patterns.
Completions stay free, but agent mode and chat draw from your credit pool at token rates. Here's how a heavy month plays out — and how much of the bill is avoidable token waste.
Illustrative example. Actual overage depends on model choice, context size and how heavily you use agent mode. The point: under token-based AI Credits, cutting tokens directly cuts dollars.
Under AI Credits, every token in agent mode and chat is billable. Terse cuts the tokens before they reach the model — right inside VS Code.
Terse's 7-stage pipeline removes filler, fixes typos, and strips redundancy from what you send Copilot Chat and agent mode. Average 40–70% fewer input tokens per request — directly stretching your AI Credit pool.
Highest ImpactAgent mode loves re-attaching the same files and re-dumping context. Terse detects redundant context before it's billed, so you don't pay token rates to send Copilot the same code it already has.
Stops Wasted TokensTerse's Agent Monitor shows the token count and credit cost of each Copilot request in real time. Catch the expensive agent turns that drain your credit pool and trigger overage — before the bill arrives.
Catch Overage EarlyCode completions and Next Edit suggestions never spend AI Credits. Terse helps you route work toward completions and keep chat/agent usage lean, so your credits go to the work that actually needs them.
Zero-Credit CodingFrontier models burn credits fastest because token rates are higher. Terse surfaces which requests are expensive so you can route simpler tasks to cheaper models and reserve premium models for hard problems.
Cheaper per TokenAI Credits count cached tokens too. By keeping context lean, Terse reduces what gets cached and re-billed across a long agent session, so a single Copilot task doesn't quietly balloon your usage.
Leaner SessionsAnswers to the most common questions about GitHub Copilot pricing in 2026.
See how GitHub Copilot stacks up against other AI coding tools — and reduce token costs across all of them.
Terse compresses prompts, flags redundant context, and tracks per-request token cost right in VS Code. 30-day free trial — no credit card required until your trial ends.