JetBrains AI has evolved – here’s what’s new
Blog|26 June 2026
JetBrains has made some of its most significant AI announcements to date in 2026, reflecting a shift in how developers are working with AI.
We’ve previously covered JetBrains’ AI innovations in our complete guide to the JetBrains AI ecosystem. But since then, there have been a number of updates made to improve the way you engage with AI.
Here’s a breakdown of the biggest changes, and why JetBrains is making them.
From AI assistant to AI agents
One of the most important developments in the story of AI and in JetBrains attitude towards it, is the evolution from simple AI assistance to fully capable coding agents.
In late 2025, JetBrains introduced Junie, its agentic coding system, which officially left beta in June 2026. Unlike traditional assistants, Junie can:
- Plan development tasks before writing code
- Edit multiple files across a project
- Run tests and fix issues
- Use the IDE’s debugger to analyse runtime behaviour
This introduces a new way of working, where tasks can truly be delegated to AI. This reflects a change where developers are increasingly expecting AI to do more than autocomplete. Instead, they want AI capable of automating entire workflows.
A move to open, multi-model AI
JetBrains is also moving away from a single-model approach. Instead, it’s embracing a flexible, open AI ecosystem.
This enables you to make use of an array of AI tools. No longer is it expected that you should hunt for one AI tool to do it all. Instead, you can bring your own API keys (including from OpenAI, Anthropic, Google), run local models for offline development, plug in external agents via the Agent Client Protocol, or make use of JetBrains-managed AI services.
This includes support for tools like Codex, Cursor, and other emerging agents directly inside the IDE.
The reason for this change is simple. The AI landscape is evolving rapidly. JetBrains recognises that the “best” AI model of today may not be the best for tomorrow, or the right fit for every job. By staying model-agnostic, they give you the freedom to choose the right tools for each task.
AI embedded across the entire development workflow
Another major shift is how deeply AI is now integrated into the JetBrains IDE experience.
With integration across the IDE, JetBrains AI is no longer limited to a chatbot or autocomplete. It’s capable of supporting with context-aware chat that understands your codebase, making file edits and refactoring, offering explanations and making improvements, and running automated tests, documentation and commit messages.
Developers don’t work in isolated snippets, they work across entire systems. JetBrains is embedding AI into real development workflows, so it adds value where it’s actually needed.
The rise of agentic workflows
JetBrains is also exploring the next step beyond single-agent workflows: multiple agents working in parallel.
Its upcoming “Air” environment allows you to:
- Run multiple AI agents simultaneously
- Assign different tasks to different agents
- Monitor and review results in one place
This introduces a model where development becomes a coordinated system of agents, rather than a one-to-one interaction with AI.
Why this change?
As AI capabilities grow, a single assistant isn’t enough. Complex development tasks benefit from specialisation and parallelisation — something multi-agent systems are designed to handle.
Supporting both AI and traditional coding
Despite these advances, JetBrains isn’t abandoning traditional development practices.
Its 2026 strategy makes it clear that the future is hybrid. The emphasis is on manual coding, while providing the ability to collaborate with AI and agents when it adds value. Both workflows coexist in the same IDE.
JetBrains recognises that not every problem should be delegated to AI.
You still want control, understanding and ownership of your code. AI is intended to enhance your workflows, not replace them.
What this means for developers
JetBrains is evolving from an IDE provider into a platform for AI-powered development. Through JetBrains, you can achieve greater automation, flexibility, and productivity, while maintaining control through hybrid workflows.
Most importantly, JetBrains is building for a future where you don’t just use AI, but collaborate with teams of AI agents, all within the tools they already trust.
And JetBrains’ new pricing model, which includes three tiers ranging from free, to premium and pay-as-you-go, offers you flexibility and choice to pick a model that suits your usage the best.
Contact Grey Matter
If you have any questions or want some extra information, complete the form below and one of the team will be in touch ASAP. If you have a specific use case, please let us know and we'll help you find the right solution faster.
By submitting this form you are agreeing to our Privacy Policy and Website Terms of Use.
Related News
Location intelligence: powering real-time decisions at scale
How do you turn disconnected real-world signals into decisions? explore how location intelligence connects data, movement and context at scale.
Developer stories: Optiyol – Optimising logistics with AI and end-to-end visibility
In this episode of Grey Matter Talks Tech, Shaun Baker is joined by Optiyol Co-Founder Ozan Gosbazi to discuss how advanced optimisation technology is reshaping logistics. They explore the shift from manual route planning to intelligent, data-driven systems that balance...
Developer stories: IP Integration – development workflows, security and growth
In this episode of our Developer Stories season on Grey Matter Talks Tech, Sam Barnes sits down with Mark Rossell from IP Integration (IPI). They explore the software IPI has built, the tools they rely on day to day, and...
Developer stories: KM Software Systems Limited – Delphi, AI and developer recruitment
In this episode of our podcast, Grey Matter Talks Tech, our Embarcadero expert Robin Ashby sits down with Eduardo Andrade, Managing Director of KM Software Systems Limited (KMSS). They dive into his long-standing career in software development and the journey...