Key findings
- One AI Credit equals US$1 of model usage, and each JetBrains AI license includes credits roughly equal to its price, with a US$5 bonus on individual AI Ultimate.
- Credits are charged at the provider's published API price per million tokens, so the model a task runs on sets its cost.
- Included credits reset every 30 days and do not roll over. Top-up credits last 12 months.
- Requests on your own provider keys or a GitHub Copilot subscription bypass credits and bill at the provider, and Junie Lite and Junie Local cost nothing.
- Organizations moved to JetBrains Central pool every license's credits and set per-user limits. Central's Business plan adds US$20 per AI seat for governance features, separate from usage.
JetBrains sells Junie through JetBrains AI licenses as a monthly model allowance in US dollars, spent faster on expensive models. Teams can move work off that allowance to their own provider keys, the free Junie Lite tier, or a model on a developer's Mac. Prices reflect JetBrains' October 12, 2026 documentation.
What each license includes
Junie and AI Assistant share a JetBrains AI license. Individual AI Pro costs US$10 monthly and AI Ultimate costs US$30. Annual billing reduces those monthly rates to US$8.33 and US$25. Organizations pay US$20 or US$60 per user per month. AI Enterprise through IDE Services costs US$60 per user per month, and JetBrains lists its credits as "on par with AI Ultimate quota, or higher". This means at least the 70 credits a month that organizational AI Ultimate includes. JetBrains says it gives no exact figure because it is "still rethinking the AI quota model", and it does not yet sell extra credits on AI Enterprise. Ask for the number in the order form.
AI Free supplies 3 credits every 30 days with an active JetBrains IDE license, including education and open source licenses. It excludes Android Studio, IntelliJ IDEA without Ultimate, and PyCharm without Pro. All Products Pack and dotUltimate include AI Pro. First activation starts a 30-day AI Pro trial.
| License | Price per user per month | AI Credits, worth US$1 each |
|---|---|---|
| AI Free | US$0 | |
| AI Pro, individual | US$10 | |
| AI Pro, organization | US$20 | |
| AI Ultimate, individual | US$30 | |
| AI Ultimate, organization | US$60 |
JetBrains AI documentation, Plans and usage, checked Oct 12, 2026, monthly billing. AI Enterprise is left out because JetBrains states its credits only as at least AI Ultimate's (rechecked Oct 25, 2026).
How credits turn into model usage
JetBrains Central prices one AI Credit at US$1 of provider usage. Its model list quotes API rates per million input, cached, and output tokens. Claude 4.6 Sonnet costs US$3 input and US$15 output per million. GPT-5.4 costs US$2.50 and US$15. Gemini 3 Flash costs US$0.50 and US$3.
GILD calculates that 35 individual AI Ultimate credits buy about 2.3 million Claude 4.6 Sonnet output tokens or 11.7 million Gemini 3 Flash output tokens. That excludes input tokens. Agents read many files before writing, so input usually dominates session cost.
- JetBrains picks the Default model on price and quality and changes it as models ship. It made Gemini 3.7 Flash the default in August 2026 at 40% off base pricing for a limited time.
- Junie also calls a helper model from the same provider for summarizing, routing, and memory, so a session draws on models the engineer did not select.
- Higher reasoning effort costs more. JetBrains recommends keeping the default effort unless a task shows it needs more.
Routes that bypass credits
Junie CLI accepts OpenAI, Anthropic, Google, xAI, OpenRouter, and GitHub Copilot credentials. When both a connected key and JetBrains AI offer a model, Junie uses the key and the provider bills directly. AI Chat accepts OpenAI or Anthropic keys. The standalone Junie plugin accepts none.
Junie Lite runs Qwen 3.6 27B free on JetBrains infrastructure in JetBrains Air and Junie CLI. JetBrains calls it "virtually unlimited" for normal use and publishes no quota, because the model behind it can change. When that model reaches its capacity, requests stop with no fallback, so keep work that must finish on credits or your own keys. Junie Local runs Qwen3.6-27B at 4-bit on an Apple Silicon M5 Mac with 64 GB. After installation it needs neither credits nor a network. Teams can also host custom models through Ollama, LM Studio, or LiteLLM without credit charges.
What organizations pay to govern it
JetBrains Central pools license credits across the organization. AI Ultimate contributes 70 each month. AI Pro, All Products Pack, and dotUltimate each contribute 20. Administrators grant access and set a default monthly limit with per-user overrides. A limit of 0 stops spending while retaining access.
Central Basic is free and includes user management, spending limits, and analytics. Business costs US$20 per AI seat per month under monthly or annual billing, excluding usage. It adds AI policies, BYOK providers, model management, MCP servers, and the Analytics API. Enterprise pricing requires a quote. Audit logs, workspaces, identity provisioning, and hybrid deployments are listed as coming soon.
Methodology
On October 12, 2026 GILD read JetBrains' licensing documentation, the JetBrains AI FAQ, the Junie home page and the Junie CLI documentation. It also read the JetBrains Central Console pages on plans, AI settings and credit limits. GILD rechecked the AI Enterprise quota and Junie Lite pages on October 25, 2026. Prices are list prices in US dollars as JetBrains states them. Local taxes and currency conversion change the final amount. Token figures are GILD calculations from JetBrains' published model price list, and exclude input and cached tokens.
Products in this piece
Sources
- Official documentation
- 15
- JetBrains AI docs: Plans and usage
- JetBrains AI FAQ
- Junie home page and pricing
- JetBrains Central Console docs: Plans and pricing
- JetBrains Central Console docs: AI settings
- JetBrains Central Console docs: AI credit limits
- Junie docs: Junie CLI quickstart
- Junie docs: Bring Your Own Key
- Junie docs: GitHub Copilot as a provider
- Junie docs: Junie Lite
- Junie docs: Junie Local
- Junie docs: Model selection
- JetBrains AI docs: Junie by JetBrains in AI Chat
- Junie docs: Plugin model settings
- JetBrains Junie blog: Junie's new default runs on Gemini 3.7 Flash (Aug 17, 2026)