Agentic AI is changing how software gets built. But for many teams, especially small and midsize ones, the path to adopting it has felt like an all-or-nothing decision: commit to a full platform subscription, or don't use AI at all.
That changes with GitLab 18.10. Starting today, Free GitLab.com teams can purchase a monthly commitment of GitLab Credits and start using GitLab Duo Agent Platform immediately. No subscription upgrade required. This is a full entry point to agentic AI for teams that aren’t ready to add a GitLab plan, but are ready to start building with AI.
The model is simple: pay for what AI does, not how many people use it. Your group owner purchases a monthly credit commitment of GitLab Credits through your group’s billing settings. Your entire team gets access to the same AI agents and flows that GitLab Premium and Ultimate customers use for planning, code generation, automated code review, and pipeline diagnosis, all drawing from a shared credit pool.
The GitLab Credits dashboard gives group owners visibility into which agents and flows are consuming credits, so you can connect AI spend directly to the work it's producing.
Day Zero with GitLab Duo Agent Platform
Once your group owner purchases credits, every member of the team can start using GitLab Duo Agent Platform right away.
Here's how a typical workflow plays out:
You start with a feature request for your software. Open Planner Agent in Agentic Chat and describe what you need in plain language. The agent breaks it down into structured work items: issues with descriptions, labels, and relationships. They are all created directly in your project. What used to be an afternoon of manual issue grooming takes minutes.
Pick up one of those issues and assign Developer Flow to begin work. The agent reads the issue context, generates code aligned to the requirements, runs tests, and opens a merge request for review. You can also use Agentic Chat for more iterative work, whether you’re refactoring, extending, or explaining code within your project's context.
When the merge request is ready, Code Review Flow runs a multi-step automated review: scanning the changes, pulling in repository context, and posting structured inline feedback tied to the diff. Your human reviewers can skip the first-pass mechanics and focus on architecture and business logic.
And if the pipeline breaks, Fix CI/CD Pipeline Flow reads the failure logs, traces the error to its root cause, and proposes a fix. Instead of manually stepping through job logs, your team gets a starting point for resolution.
GitLab’s Duo Agent Platform takes software development from iteration to deployment, powered by one credit pool.
It's simple to get started with agents and flows, from plan to deploy in under 3 minutes. Watch this demo to learn how:
Flat-rate code review: Predictable costs as you scale
Of all the workflows available through GitLab Duo Agent Platform, automated code review is where teams see value the fastest, and where predictable pricing matters most.
Code Review Flow now costs a flat 0.25 GitLab Credits per review, regardless of merge request size, repository complexity, or how many steps the flow runs internally. Four reviews equal one credit. Whether your team merges 500 requests a month or 50,000, you can forecast costs directly in terms of reviews.
The math is worth a closer look. Manual code reviews don’t just cost money, they take time and add disruption in development with constant context-switching. The time saved with the Code Review Flow could mean substantial savings as review volume grows. Now you have the potential to run hundreds of reviews simultaneously rather than having them sit in a queue. That means the time savings quickly compound with the cost savings.
For teams on GitLab’s Free tier, this means you know exactly what portion of your monthly credit pool goes to code review and can plan accordingly.
Learn more about how Code Review Flow works and what it means for scaling your engineering organization.
Why Premium multiplies the value
GitLab Credits on the Free tier give your team a direct path to agentic AI. If your team relies on GitLab for more, Premium is where the economics and the capabilities come together.
At $29 per user per month, GitLab Premium includes 12 GitLab Credits per user as a promotional offer. For a team of 20, that's 240 credits per month before you spend anything extra, enough to cover roughly 960 automated code reviews, or a mix of code review, planning, development flows, and pipeline fixes.
GitLab Duo Agent Platform is only a part of what Premium gives you. You also get advanced CI/CD for high-volume pipelines, merge approvals and code owners for governance, and AI that runs within a single data layer with unified context across your projects.
If your team is spending credits on Free and finding that AI is becoming central to your workflow, Premium is the natural next step with the included promotional credits. It offers more platform capability, and a foundation that scales with you.
Get started today
GitLab 18.10 is available now. Whether your team needs agentic AI to move faster or the full platform to support how you already work, there's a clear path for accelerating your software development process.
- Free GitLab.com teams: Purchase a monthly commitment of GitLab Credits through your group billing settings and start using GitLab Duo Agent Platform today.
- Teams ready for the full platform: Find the right GitLab subscription for your team, or start a free trial of GitLab Ultimate.
Getting credits set up for your team is fast and easy. Watch this demo to learn how:
FAQ
What is a monthly commitment of GitLab Credits?
A monthly commitment is a usage-based purchase option where your group owner selects a set number of credits that apply as a shared pool across the group. Credits are consumed when your team uses GitLab Duo Agent Platform capabilities. Learn more in the GitLab Credits documentation.
Who can purchase GitLab Credits today?
GitLab Premium and Ultimate customers already receive promotional credits included with their subscription. Starting in 18.10, Free GitLab.com top-level group namespaces can also purchase a monthly commitment of credits through self-serve group billing. For the latest on eligibility, see the GitLab Credits documentation.
What AI capabilities do credits unlock on Free?
Teams with credits get access to the same agentic AI capabilities and models available to Premium and Ultimate customers, including Planner Agent, Developer Flow, Code Review Flow, Fix CI/CD Pipeline Flow, Agentic Chat, Code Suggestions, custom agents and flows, and more. For the full list, see the Duo Agent Platform documentation.
How much does automated code review cost?
Code Review Flow charges a flat rate of 0.25 GitLab Credits per review, regardless of merge request size or complexity. For current pricing details, see the Code Review Flow documentation.
Can I upgrade from Free with credits to GitLab Premium?
In GitLab 18.10, upgrading from a Free namespace with a monthly credit commitment to Premium is available through a sales-assisted path. Contact the GitLab sales team to explore your options.
Facts Only
GitLab 18.10 is now available.
Free GitLab.com teams can purchase monthly GitLab Credits to access GitLab Duo Agent Platform.
No subscription upgrade is required for Free tier teams to use AI capabilities.
GitLab Credits are purchased through group billing settings and shared across the team.
AI capabilities include Planner Agent, Developer Flow, Code Review Flow, Fix CI/CD Pipeline Flow, and Agentic Chat.
Automated code review costs a flat 0.25 GitLab Credits per review.
GitLab Premium costs $29 per user per month and includes 12 promotional credits per user.
Premium includes additional features like advanced CI/CD, merge approvals, and unified AI context.
Free tier teams can upgrade to Premium via a sales-assisted path.
GitLab Credits dashboard provides visibility into credit consumption.
The update aims to make agentic AI accessible to smaller teams without requiring a full platform commitment.
GitLab Ultimate is available for a free trial.
Executive Summary
GitLab 18.10 introduces a new way for Free GitLab.com teams to access agentic AI capabilities through GitLab Credits, eliminating the need for a full platform subscription. Teams can now purchase a monthly commitment of credits, which are shared across the group and consumed by AI-driven workflows like planning, code generation, automated code review, and pipeline diagnosis. The model is usage-based, charging for AI actions rather than per-user access. For example, automated code reviews cost a flat 0.25 credits per review, regardless of complexity, making costs predictable as teams scale. Free tier users gain access to the same AI agents and flows as Premium and Ultimate customers, including Planner Agent, Developer Flow, and Code Review Flow. GitLab Premium, priced at $29 per user per month, includes 12 promotional credits per user, offering additional platform capabilities like advanced CI/CD and governance tools. The update aims to lower the barrier to AI adoption for smaller teams while providing a clear path to scaling with Premium features.
The announcement highlights practical workflows, such as using AI to break down feature requests into structured issues, generate code, and automate merge request reviews, reducing manual effort and accelerating development cycles. GitLab emphasizes transparency with a credits dashboard to track usage and costs. While Free tier teams can now experiment with AI, the promotional credits in Premium suggest a strategic push to convert users to higher-tier plans as their reliance on AI grows. The move reflects a broader trend in software development tools toward flexible, consumption-based pricing models for AI features.
Full Take
**STEELMAN:** GitLab’s new credit-based model for AI access is a pragmatic response to the adoption barriers faced by smaller teams. By decoupling AI usage from subscription tiers, it democratizes access to powerful tools like automated code review and planning agents, which can significantly reduce manual labor and accelerate development. The flat-rate pricing for code reviews, in particular, addresses a common pain point—predictable costs at scale—while the transparency of the credits dashboard empowers teams to manage spending effectively. This approach aligns with broader industry shifts toward usage-based pricing, making advanced tooling more accessible without forcing teams into rigid licensing structures.
**PATTERN SCAN:** The narrative leans heavily on the appeal of flexibility and cost predictability, which are legitimate advantages. However, the framing of "no subscription upgrade required" subtly implies that the current model is more liberating than it may be—teams still need to commit to monthly credit purchases, and the path to Premium is positioned as a natural progression. This could be seen as a form of **ARC-0024 Ambiguity**, where the emphasis on "no upgrade" obscures the fact that teams are still entering a financial commitment. Additionally, the repeated contrast between Free and Premium tiers might subtly pressure teams to view the latter as the inevitable next step, a soft **ARC-0043 Motte-and-Bailey** where the "motte" is accessibility and the "bailey" is upselling.
**ROOT CAUSE:** The paradigm here is the commodification of AI in software development, where access is gated not by technical barriers but by pricing models. The unstated assumption is that AI-driven workflows are inherently more efficient, which may not account for the learning curve or the potential for over-reliance on automation. Historically, this echoes the SaaS industry’s shift from perpetual licenses to subscription models, now evolving into microtransactions for AI features.
**IMPLICATIONS:** For human agency, this model could empower smaller teams to compete with larger ones by leveraging AI, but it also risks creating dependency on GitLab’s ecosystem. The cost savings from automated reviews might be offset by the need to purchase more credits as usage grows, and teams could find themselves locked into a platform where switching becomes costly. The second-order consequence is the potential homogenization of development workflows, as teams optimize for AI-assisted processes rather than diverse, human-driven approaches.
**BRIDGE QUESTIONS:**
How might the credit-based model change if AI tools become commoditized across platforms?
What are the long-term costs of dependency on proprietary AI agents for core development tasks?
How does this model compare to open-source alternatives in terms of flexibility and control?
**COUNTERSTRIKE SCAN:** If this were part of a coordinated influence campaign, the playbook would emphasize "democratization" and "flexibility" to frame GitLab as a benevolent disruptor, while subtly steering teams toward higher-tier plans. The actual content aligns with this pattern but stops short of overt manipulation—it’s a legitimate business strategy, not a deceptive one. The focus on transparency (e.g., the credits dashboard) mitigates concerns about hidden costs, though the long-term economic incentives for GitLab remain clear.
**Patterns detected: ARC-0024 Ambiguity, ARC-0043 Motte-and-Bailey**
Sentinel — Human
The article shows strong signs of human authorship, with natural variability in style and product-specific insights, though some sections exhibit mild stylometric uniformity.