OpenAI Codex Lands in JetBrains IDEs with Free Credits

OpenAI Codex has arrived in JetBrains IDEs. Powered by GPT-5.2-Codex, this autonomous coding agent can now debug, refactor, and build features directly within IntelliJ, PyCharm, WebStorm, and Rider.

JetBrains announced free promotional access to Codex through JetBrains AI subscriptions starting January 22, with credits available until the allocated pool is exhausted. Integration requires version 2025.3 or later of any supported JetBrains IDE.

“By partnering with leading AI providers like OpenAI and integrating their technologies directly into JetBrains IDEs, we’re ensuring these tools work where developers already are, in a way that respects their preferences,” — JetBrains, official announcement.

With this partnership, JetBrains positions itself to compete more directly against Microsoft’s VS Code ecosystem, which has enjoyed exclusive access to advanced AI coding tools. By offering promotional credits, JetBrains is betting that developers who try Codex will find enough value to convert to paying subscribers once the free tier expires.

What Makes Codex Different

Codex represents a fundamental shift from traditional code completion tools. When OpenAI reintroduced Codex as an autonomous software engineering agent in 2025, it marked a departure from the line-by-line suggestion model that powered early versions of GitHub Copilot.

Modern Codex can read and write files, run tests until they pass, and submit pull requests autonomously. Originally, Codex powered the first version of GitHub Copilot before Microsoft’s tool migrated to newer OpenAI models.

Such evolution signals a broader industry trend toward agentic development, where AI handles entire workflows rather than individual code snippets. As a result, development teams may need to rethink how they allocate work, with human developers focusing on architecture and review while AI agents handle implementation details.

Rather than reviewing suggestions line by line, developers using Codex can delegate entire tasks and review completed work. In practice, code review could shift from catching AI mistakes to validating AI decisions.

To accommodate varying comfort levels with AI autonomy, Codex supports Chat, Agent, and Agent (Full Access) approval modes. A Cloud Delegation feature allows offloading longer jobs to a cloud environment, letting developers continue working locally while Codex handles complex tasks in the background. Through this tiered approach, developers can gradually trust AI with increasingly complex tasks, starting with suggestions and progressing to fully autonomous operations.

Technical Details and Access

Beyond its autonomous capabilities, Codex offers flexible authentication options. Developers can access it through three different methods: JetBrains AI subscription, ChatGPT account, or OpenAI API key (BYOK).

Free promotional access only applies when using the JetBrains AI subscription method. ChatGPT Plus, Pro, Business, Edu, and Enterprise plans include Codex access for those who prefer OpenAI’s direct authentication. After the promotional period ends, Codex usage through JetBrains AI will consume AI Credits. JetBrains reserves the right to cancel the promotion at any time without prior notice.

Users can also switch between supported OpenAI models and adjust reasoning budget directly in the AI chat interface. Reasoning effort can be set to low, medium, or high, allowing developers to trade off response speed against depth of analysis depending on task complexity.

JetBrains AI Assistant also connects to different LLMs from providers including Google, plus local models for developers who need offline work capabilities. For enterprises already invested in OpenAI’s ecosystem through ChatGPT Business or Enterprise plans, integration adds value to existing subscriptions rather than requiring new purchases.

Advanced Capabilities

Building on these foundations, GPT-5.2-Codex brings enhanced security analysis capabilities to JetBrains users. Released in December 2025 with context compaction and Windows software optimization features, the model can handle large refactors, code migrations, and feature builds spanning multiple files.

For security-conscious enterprises, automated vulnerability detection during the coding phase addresses a gap that typically requires separate static analysis tools.

What emerges from this announcement is that OpenAI now competes on several fronts: powering Microsoft’s GitHub Copilot while simultaneously offering a direct alternative through JetBrains. Such a dual strategy could accelerate feature development as both platforms vie for developer attention.

JetBrains’ announcement follows OpenAI’s August 2025 Codex IDE integration for VS Code and the tool’s general availability launch in October 2025. Developers interested in testing Codex should act during the promotional window, as JetBrains has not specified when free credits will run out.