GitHub Copilot

A coding AI that autocompletes and writes code inside dev tools such as VS Code

Coding

CurrencyUSDKRWEUR

Personal starting price$10+Official price in USD
Free tierFree plan available
Korean supportGood

Price checked 2026-10-17, source: github.com/features/copilot/plans. Prices and benefits may change. Check the official site before you pay.

Pros

  • Widest integration with dev tools
  • Can choose models from several companies
  • Free for students and open source

Cons

  • Usage-based AI credits run out fast if you use top models a lot
  • May be weaker than dedicated AI editors for complex tasks

These are opinions compiled by Pickx from public sources.

At a glance

GitHub Copilot is a coding AI that autocompletes code in VS Code, Visual Studio, JetBrains, the terminal (Copilot CLI) and GitHub.com, fixes code in chat and agent modes, and even creates a PR when you hand it an issue. You can choose models from several companies, such as OpenAI, Anthropic, Google, xAI, Microsoft and Moonshot. From June 1, 2026, billing changed from 'premium requests' to GitHub AI credits (1 credit = $0.01, deducted by each model's token usage), and code autocompletion remains unlimited on paid plans.

Full model list

The Free and Student plans cannot choose a model directly and only get automatic selection (Auto). Paid plans can choose from the models below, and credits are deducted at each model's API rate (prices are per 1,000,000 tokens, input/output).

ModelTypeFeaturesAvailable plans
GPT-6.1 Sol (OpenAI)Coding/reasoning
  • OpenAI's latest Sol model
  • Strong at writing and refactoring complex code
  • Good performance for the price
Paid plans (Pro and above)
GPT-6 Astra (OpenAI)Coding/reasoning (top tier)
  • OpenAI's highest-performance model
  • For hard design work and large-scale changes
  • Uses a lot of credits
Paid plans (Pro+/Max/Business and above recommended)
GPT-6 Sol (OpenAI)Coding/reasoning
  • Main model of the GPT-6 generation
  • Balanced speed and performance
Paid plans
GPT-6 Luna (OpenAI)Coding (lightweight)
  • Very cheap, fast lightweight model
  • Suited to simple fixes and explanations
Paid plans
GPT-5.6 Sol (OpenAI, previous model)Coding/reasoningPrevious-generation flagshipPaid plans
GPT-5.6 Terra (OpenAI, previous model)CodingPrevious-generation mid-range modelPaid plans
GPT-5.6 Luna (OpenAI, previous model)Coding (lightweight)Previous-generation lightweight modelPaid plans
GPT-5.5 (OpenAI, previous model)Coding/reasoning
  • $5/$30 (272K or less)
  • Stable general-purpose model
Paid plans
GPT-5.4 (OpenAI, previous model)Coding/reasoning
  • March 2026 model
  • Solid for agent work
Paid plans
GPT-5.4 mini (OpenAI)Coding (lightweight)Cheap small modelPaid plans
GPT-5.4 nano (OpenAI)Coding (ultra-lightweight)
  • Used only in the Codex VS Code extension
  • Not available in Copilot chat
Pro+ only
GPT-5.3-Codex (OpenAI)Coding-specialized
  • Coding-only model of the Codex family
  • Strong at long agent tasks
Paid plans
GPT-5 mini (OpenAI)Coding (lightweight)
  • $0.25/$2, among the cheapest
  • For light questions
Paid plans (also used in automatic selection)
Claude Opus 5.5 (Anthropic)Coding/reasoning (top tier)
  • $4/$20
  • Top-tier for complex multi-file work and agentic coding
Paid plans (Pro+/Max/Business and above recommended)
Claude Opus 5 (Anthropic)Coding/reasoningPrevious Opus modelPaid plans
Claude Opus 4.8 (Anthropic, previous model)Coding/reasoning
  • Previous-generation Opus
  • Has a fast mode version (preview)
Paid plans
Claude Sonnet 5.5 (Anthropic)Coding
  • Good balance of speed and quality
  • Widely used as the everyday coding workhorse
Paid plans
Claude Sonnet 5 (Anthropic)CodingPrevious Sonnet modelPaid plans
Claude Sonnet 4.6 (Anthropic, previous model)Coding
  • $3/$15
  • Previous-generation Sonnet
Paid plans
Claude Haiku 5.5 (Anthropic)Coding (lightweight)Fast, cheap small modelPaid plans
Claude Haiku 4.5 (Anthropic, previous model)Coding (lightweight)Previous-generation small modelPaid plans
  • Claude Fable 5
  • Fable 5.1 (Anthropic)
Coding/reasoning
  • Disabled by default for companies
  • Requires admin approval, a zero data retention agreement and the like
Paid plans (companies need admin activation)
Gemini 3.8 Flash (Google)Coding (fast)
  • $0.75/$3.75 (promotion until the end of 2026)
  • Fast
  • Strong at long-context handling
Paid plans
Gemini 3.7 Flash (Google)Coding (fast)Previous Flash modelPaid plans
Grok 4.7 (xAI)Coding/reasoning
  • $2/$6 (200K or less)
  • Context 500,000 tokens
  • On the cheap side
Paid plans
  • Grok 4.6
  • Grok 4.5 (xAI, previous model)
Coding/reasoningPrevious Grok modelsPaid plans
MAI-Code-1.1-Flash (Microsoft)Coding (lightweight)
  • $0.20/$1.20
  • Microsoft's own coding model released 2026-10-07
  • Fast and cheap
Paid plans
Kimi K3 (Moonshot AI)Coding (open weights)
  • $3/$15
  • Open-weights model
  • Disabled by default for companies
Paid plans
Auto (automatic selection)AutomaticCopilot automatically picks the model that fits the taskAll plans (Free/Student get only this)

Recently retired models: Claude Opus 4.7, Gemini 3.5/3.6 Flash, Kimi K2.7 Code (2026-10-02), MAI-Code-1-Flash (9-10), Claude Opus 4.5/4.6, Sonnet 4.5, Gemini 3.1 Pro (9-01) and others. The plan restrictions per model change often in the official docs table, so check the model picker in your IDE for the final answer.

Summary by plan

PlanPriceWhat you can doUsage limitsFeatures
Copilot Free$0
  • Code autocompletion
  • Chat
  • Agent mode
  • MCP
  • Copilot CLI
  • Models by automatic selection only
  • Code autocompletion 2,000 per month
  • Limited AI credits at about 50 chats per month
  • No coding agent (issue to PR) or code review
  • Only for individuals without company Copilot
Copilot Student$0 (student verification)
  • Automatic model selection only
  • Delegation to third-party agents excluded
Limited AI credits (figures not published officially)Verified students only
Copilot Pro
  • $10/mo
  • Annual billing not offered to new subscribers (needs checking)
  • Coding agent
  • Code review
  • Java/.NET app modernization
  • Choose models directly
  • Extra credits can be purchased
  • 1,500 AI credits per month (worth $15 = 1,000 base + 500 extra)
  • Unlimited autocompletion
Teachers and popular open-source maintainers may get it free
Copilot Pro+$39/mo
  • Access to all premium models
  • Delegate work to third-party agents such as Claude and OpenAI Codex (preview)
  • GPT-5.4 nano (Codex extension)
7,000 AI credits per month (worth $70 = 3,900 base + 3,100 extra)GPT-5.4 nano is Pro+ only
Copilot Max$100/mo
  • Priority access to premium models
  • All Pro+ features
20,000 AI credits per month (worth $200 = 10,000 base + 10,000 extra)New top personal plan created in 2026
Copilot Business$19 per user per month
  • Organization license and policy management
  • IP indemnity
  • Content exclusion
  • Audit logs
  • Admin budget settings
  • 1,900 credits per user per month (worth $19, shared across the organization)
  • Overage $0.01 per credit
Team plan with credit pooling
Copilot Enterprise$39 per user per monthPriority access to premium models / GitHub.com chat / Codebase indexing / Enterprise features such as custom models3,900 credits per user per month (worth $39) pooledRequires GitHub Enterprise Cloud

From 2026-06-01, billing changed to usage-based (AI credits). Existing Pro/Pro+ annual subscribers stay on the old 'premium requests' system until expiry and switch to Free when it expires (if you switch to monthly billing, the remaining amount is converted to credits). There is no separate official Korean won price (paid in dollars).

Which plan for whom

  • Copilot Free: People who are just starting to code or occasionally do side projects. You can try autocompletion and agent mode without commitment.
  • Copilot Student: If you can verify as a student, apply for this first without question.
  • Copilot Pro: Developers who code every day but mainly use autocompletion and light chat. At $10/mo with unlimited autocompletion, the value is best.
  • Copilot Pro+: Developers who often run agent work with top models like Claude Opus/GPT-6 Astra, and people who want to use third-party coding agents in one place too.
  • Copilot Max: Heavy users and freelancers who run several coding agents all day. Consider it if Pro+ credits run short every month.
  • Copilot Business: Companies that adopt it by development team and need to manage policy, security and budget.
  • Copilot Enterprise: Large companies using GitHub Enterprise Cloud, and organizations that want tailored answers by indexing their internal codebase.

Notes

Good for

#AI#Artificial Intelligence#Generative AI#Coding#Development#Study#Free

7 views (2 today)

Prices converted at 1 USD = ₩1,329.49 (2026-10-19). Original price shown in brackets. Taxes as listed by each provider.

Comments

No comments yet. Be the first.

Your comments and nickname are public. If you see abuse or advertising, report it here and we will review and remove it.