GitHub Copilotの料金まとめ【2026年7月】個人4プラン化とAIクレジット制をわかりやすく整理GitHub Copilot's individual plans were restructured into four tiers—Free, Pro…

GitHub Copilot's pricing has undergone a significant restructuring in 2026, and for anyone weighing the cost of an AI coding assistant, understanding the new shape of the plans matters as much as the tools themselves. As of late July 2026, GitHub's individual subscriptions are organized into four tiers, and a new "AI credit" model that began on June 1, 2026 now sits underneath them. This piece summarizes the current plan structure and how the credit system works, based on GitHub's official pricing displayed in US dollars, with yen figures given only as a rough guide at an assumed rate of 1 USD to 150 yen.

The headline change is that individual plans now span four levels rather than the simpler split of earlier years. At the bottom is Free, which carries no monthly charge. Above it sits Pro at 10 dollars per month, followed by Pro+ at 39 dollars per month, and finally Max at 100 dollars per month. Roughly translated at the stated rate, that places Pro near 1,500 yen, Pro+ around 5,850 yen, and Max at about 15,000 yen monthly. Because GitHub sets prices in dollars, the actual amount charged in other regions can shift with exchange rates and local billing, so the official site remains the authoritative reference for exact figures.

The tiered structure appears designed to separate casual users from heavy users more clearly than before. A free entry point lowers the barrier for developers who want to try Copilot's completions and chat features, while the paid steps up to Max are aimed at those who lean on the assistant throughout the working day. The wide gap between the 10 dollar Pro plan and the 100 dollar Max plan suggests GitHub is targeting distinct segments, from individual hobbyists to professionals who depend on the most capable models and the largest usage allowances.

The second major shift is the AI credit system introduced at the start of June 2026. Under this model, usage of certain premium capabilities is metered through credits rather than being treated as effectively unlimited. Each plan appears to include an allotment of credits, with higher tiers granting larger amounts, and the intent is likely to tie the cost of running more expensive model requests to the price of the plan. This kind of consumption-based layer has become common across the AI industry as providers try to align pricing with the real compute cost of large language models. Because specifics such as credit quantities and per-action consumption can change, users should confirm the current allowances directly through GitHub before committing to a plan.

For context, the move mirrors a broader trend among AI developer tools. Copilot competes with assistants such as Amazon's Q Developer, Google's Gemini Code Assist, Cursor, and a range of editor integrations built on models from providers like OpenAI and Anthropic. Many of these products have shifted toward multi-tier pricing and metered access as they expose more powerful, and more costly, models. The credit approach lets a vendor offer premium reasoning models without exposing itself to unbounded expense from a small number of extremely heavy users. It also gives customers a clearer signal about which features are cheap and which are resource intensive.

It is worth remembering that Copilot began as a single autocomplete-style tool built on OpenAI's Codex and has since expanded into chat, agent-style workflows, code review assistance, and integrations across the GitHub platform and popular editors. The layered pricing reflects that widening scope: as the assistant does more, GitHub has more reasons to differentiate what each subscription level unlocks. Enterprise and business plans exist separately from these individual tiers, so organizations evaluating Copilot at scale should treat the four consumer plans as only part of the overall lineup.

For individual developers, the practical takeaway is to match a plan to actual usage rather than defaulting to the cheapest or the most expensive option. Someone experimenting occasionally may find Free or Pro sufficient, while a developer relying on advanced models and higher credit limits throughout the day may justify Pro+ or Max. Given that both the prices and the credit rules were revised only recently and are quoted in dollars, checking GitHub's official pricing page for the latest details remains the safest way to avoid surprises on a monthly bill.