Amazon Q Developer is on its way out. AWS blocked new Free Tier accounts and new Pro subscriptions on May 15, 2026, removed Opus 4.6 from Q Developer Pro on May 29, and set full end of support for the IDE plugins and paid subscriptions at April 30, 2027. The replacement is Kiro, an agentic spec-driven IDE built on Code OSS, where the newest models now live exclusively. Existing Pro subscriptions can still add users until the 2027 deadline, and the AWS Console version plus Slack and Teams integrations are unaffected.
Reviewed by EdGrows·Updated Aug 18, 2026Affiliate link — we may earn a commission
This analysis is based on documentation, public user reports, and vendor materials — not yet on our own hands-on testing. How we rate
If you use Amazon Q Developer, four dates decide your next year:
That's twelve months of runway from the announcement, which is a more generous transition than several vendors managed in 2026.
Two things are not being switched off: the AWS Management Console version of Q Developer, and the Slack and Teams integrations. If that's where you use it, nothing changes for you.
Kiro is not a renamed Q Developer. It's a different idea about how AI should sit in development.
Q Developer worked the way most assistants do: you ask, it suggests, you accept or reject. Kiro requires a structured specification before code generation begins — AWS calls it spec-driven development, and the IDE is deliberately built so that skipping the spec is awkward rather than merely discouraged.
The mechanics: built on Code OSS, so VS Code settings and Open VSX extensions carry over. Models run on Amazon Bedrock. Pricing starts free at around 50 agentic interactions a month, with Pro near $19–20 for about 1,000 interactions, and a credit system that shows what each prompt costs as you work.
That per-prompt visibility is genuinely uncommon. Most subscriptions hide consumption until you hit a limit; Kiro puts the meter in front of you.
AWS's reasoning is that reactive autocomplete fails at production scale, and that specifying before generating catches errors that patching afterward never will.
There's something to that, and it's also a bet. Spec-driven development asks for discipline up front, which is exactly the thing developers skip when a deadline is close. Whether teams sustain it past week three is the open question, and nobody has a year of evidence yet.
The forced-migration framing matters too. Tens of thousands of enterprise teams adopted Q Developer over two years and are now being moved to a tool with a different workflow, not a newer version of the same one. That's a reasonable moment to evaluate the whole category rather than accept the default.
If you're on the free tier: you have no subscription to protect. Start trialling replacements now — there's no cost to comparing Kiro against GitHub Copilot, Cursor and Claude Code properly.
If you're on Pro: you can keep adding seats until 2027, which removes any urgency. Use that runway. Run Kiro alongside your existing setup on a real project before committing an organisation to spec-driven development.
If you depend on code transformation: this needs its own plan. Java 8/11/17 to 17/21 upgrades and .NET porting run through the plugins and disappear with them. A modernisation programme built on that capability should be scoped and scheduled now, not in Q1 2027.
If you use the Console or Slack integrations: nothing to do.
Worth recording, because it explains who should look hardest at Kiro rather than at general-purpose alternatives.
It knew AWS. Answering account-level questions, generating CLI commands, explaining IAM policies and CloudFormation templates, reasoning about Lambda and service configuration — a general assistant does that from documentation, Q Developer did it from inside the platform. It also carried compliance eligibility for regulated environments, which is frequently the gating factor before any AI tool reaches a developer's machine.
Where it never competed was general coding quality. Against GitHub Copilot on everyday completion, or Cursor on multi-file work, it was adequate rather than preferred. CodeWhisperer before it struggled to gain traction for the same reason.
That gap is the honest explanation for why AWS rebuilt instead of iterating.
Against Kiro: the intended path, and the only one that keeps AWS-native depth plus the newest models. If your work is infrastructure-heavy — HCL, CloudFormation, Lambda — this is the shortest migration. The cost is adopting a spec-first workflow that isn't optional.
Against GitHub Copilot: Copilot is $10 with unmetered inline completion, works across VS Code, JetBrains, Visual Studio and Neovim, and has the largest install base in the category. It knows less about AWS and more about code. For most teams leaving Q Developer, this is the safe default.
Against Cursor: Cursor at $20 has the most mature agent workflows and a full editor built around them. It's the strongest general-purpose option and knows nothing special about your AWS account.
Against Claude Code: Claude Code runs in the terminal, suits delegated multi-file work, and pairs naturally with infrastructure workflows. Similar price to Kiro Pro, no spec requirement.
Against Aider: Aider is free and model-agnostic, which is worth knowing for anyone whose lesson from this sunset is that vendor-controlled tools can be taken away. Apache 2.0 code on your disk doesn't get end-of-lifed.
| Item | Status |
|---|---|
| Q Developer Free Tier | Closed to new accounts since May 15, 2026 |
| Q Developer Pro ($19/user/mo) | Closed to new subscriptions; existing can add seats |
| Q Developer IDE plugins | End of support April 30, 2027 |
| AWS Console Q Developer | Unaffected |
| Slack / Teams integrations | Unaffected |
| Kiro Free | ~50 agentic interactions/month |
| Kiro Pro | ~$19–20/month, ~1,000 interactions |
Checked August 2026 against AWS's own end-of-support announcement and current reporting. New Q Developer signups closed May 15, 2026; Opus 4.6 was removed from Q Developer Pro on May 29; full end of support for IDE plugins and paid subscriptions is April 30, 2027. Kiro pricing is reported around $19–20 for the Pro tier — confirm on AWS's Kiro pages, since the credit allowances are newer than the price point and more likely to move.
Don't wait for the deadline to evaluate. Twelve months sounds long and disappears quickly when it involves changing how a team writes code.
Trial Kiro on a real project, not a demo. Spec-driven development either fits how your team works or it doesn't, and a toy example won't tell you which.
Scope the code-transformation dependency separately. It's the one capability that has no obvious equivalent elsewhere.
Treat this as a category review. You're being moved regardless — that's the cheapest moment to compare Copilot, Cursor and Claude Code properly.
Amazon Q Developer is being retired, and the honest job of this page is to say so clearly rather than review a product you can no longer buy. New signups closed on May 15, 2026, the newest models moved to Kiro two weeks later, and the IDE plugins and paid subscriptions end support on April 30, 2027.
The product's strengths were real and narrow. Deep AWS knowledge — account-level questions, CLI generation, IAM and CloudFormation reasoning — plus compliance eligibility for regulated environments, which is often what determines whether an AI tool reaches developers at all. Its weakness was equally consistent: on general coding it never matched Copilot or Cursor, and adoption reflected that.
Kiro is a credible replacement and a genuine change of workflow rather than a version bump. Spec-driven development asks for discipline before generation, which is a defensible response to agents writing more code than teams can review — and an unproven one, since the industry has roughly a quarter of evidence on whether teams sustain it.
For existing Pro customers, the runway is generous and should be spent evaluating rather than migrating on autopilot. For anyone choosing a coding assistant today, this isn't an option — the signup door is shut.
Note: Amazon Q Developer does not have an affiliate program with AIVario. We earn no commission, and this rating reflects the product's discontinued status rather than the quality of its AWS integration.
Best for: Existing Q Developer teams planning migration before April 2027; AWS shops evaluating whether Kiro's spec-driven model fits them Not ideal for: New adoption of any kind, teams needing best-in-class general coding, organisations depending on code transformation past 2027 Bottom line: Closed to new signups, support ends April 30, 2027, and the newest models already live in Kiro. Use the runway to evaluate the whole category, not just the successor.
No. AWS blocked new Free Tier account creation via Builder ID and new Pro subscription creation via the AWS Console on May 15, 2026. The one exception is seat expansion: if your organisation already holds an active Q Developer Pro subscription, you can keep adding users to it until the end-of-support date. Everyone else evaluating AWS coding tools is being pointed at Kiro.
April 30, 2027 for the IDE plugins and paid subscriptions, which gave customers twelve months' notice from the announcement. Critical bugfixes continue to ship to existing users through the transition window, and the plugin listings stay published on all four IDE marketplaces with a deprecation notice pointing at Kiro. The AWS Management Console version of Q Developer and the Slack and Teams integrations are not part of this sunset.
AWS's replacement — an agentic IDE built on Code OSS, so VS Code settings and Open VSX extensions carry over. Its distinguishing idea is spec-driven development: rather than prompting and iterating toward working code, Kiro requires a structured specification before generation begins, and the IDE deliberately makes skipping that step difficult. Models run on Amazon Bedrock, and it's where AWS now ships its newest Claude access.
AWS moved its newest model access to Kiro to make the migration path obvious. Opus 4.6 stopped being available on Q Developer Pro on May 29, 2026, with Opus 4.5 and the other existing models staying put. The latest coding models, including Opus 4.7, are available only through Kiro. If you're on Q Developer Pro and wondering why your model list looks dated, that's why.
It starts free with around 50 agentic interactions a month, with Pro at roughly $19–20 a month for about 1,000 interactions and higher tiers above that. The credit model gives per-prompt visibility into what each task consumes, which is unusual — most flat-rate subscriptions keep usage opaque until you hit a wall. At that price it sits alongside Cursor Pro and Claude Code Pro, both also around $20.
No, and it's worth treating the deadline as a genuine evaluation rather than an automatic migration. Kiro is the path of least resistance for AWS-committed teams, especially for infrastructure work where HCL and CloudFormation support matter. But GitHub Copilot, Cursor and Claude Code are all mature options, and the spec-driven workflow is a real change in how you work rather than a like-for-like replacement. Twelve months is enough time to trial two or three properly.
It's included in the end of support. Java 8/11/17 to 17/21 upgrades and .NET porting are delivered through the Q Developer IDE plugins and require a paid subscription, so they go away with the plugins on April 30, 2027. Any organisation with a modernisation programme running on that capability needs to plan for it specifically rather than assume the feature migrates.
Solid inside AWS, unremarkable outside it. It knew AWS services deeply — answering account-level questions, generating CLI commands, explaining IAM and CloudFormation — and carried the compliance eligibility that regulated AWS shops need. What it never did was win on general coding quality against Copilot or Cursor, which is ultimately why AWS rebuilt rather than iterated. Adoption never matched CodeWhisperer's ambitions either.
The fastest issue tracker for engineering teams. Used by Vercel, Loom, Retool, and thousands of modern startups. Free up to 250 issues.
The AI code editor that became the default — now rebuilt around parallel agents in Cursor 3, with its own Composer models.
Terminal-native, model-agnostic, git-obsessed. Cheaper than the vendor CLIs and slightly less accurate — an explicit trade, not a hidden one.