Flutter Q2 2026 Survey: Trust at 83%, Google at 62%, and an Honesty Gap
3,500+ developers answered Flutter's Q2 2026 survey. Satisfaction held at 93%, trust in Flutter hit 83%, and the team published where it's falling short.
Published on • October 19, 2026
AI Assistant
The Flutter Q2 2026 survey results, published August 6, 2026 by Emma Twersky, are notable less for the headline numbers than for what the team did with them: it reviewed recent releases and the 2026 roadmap against the feedback and published where the alignment doesn’t hold — “without sugar-coating.” The survey ran June 8–22, collected over 3,500 complete responses through the IDE plugins (VS Code, IntelliJ, Android Studio), X, and the website, and has run twice a year since Flutter 1.0 in 2018.
Satisfaction held at 93% — but the mix shifted
93% of respondents reported positive satisfaction (top-2 box: very + somewhat), steady against Q4 2025. Underneath, the distribution moved: 58% are now very satisfied, a statistically significant 6-point jump from 52%, while “somewhat satisfied” fell from 40% to 35%. Trust in Flutter to consistently meet development needs moved the same direction, rising from 77% to 83%.
The survey tracks 16 focus areas because a single score hides variance. This cycle, every measured developer subgroup — by experience, company size, team size, and platform — reached 90% or higher, a first in survey history. The highest-rated areas: Dart language at 92%, Android app development at 91%, and the core framework at 90%.
Novice developers, historically the least-satisfied segment, hit 90% top-2 box (up 5 points), which the team connects to the guided Getting Started pathway launched in February. But isolate the core framework for that same group and an 11-point gap appears: 82% for novices versus 93% for experts. The team’s read is blunt — onboarding tutorials alone can’t solve a steep core-framework learning curve. One likely driver it names: novices report lower AI-tool adoption, and the team is now exploring surfacing AI-assisted Flutter guidance earlier.
AI agents are already default infrastructure
The survey added editor and agent usage for the first time. VS Code leads at 66% and Android Studio at 40% — but Claude Code (32%) and Antigravity (23%) now sit ahead of GitHub Copilot (19%), Cursor (18%), and Codex (17%). (Respondents could pick multiple tools, so these sum past 100%.) Satisfaction clustered high: VS Code 88%, Claude Code 86%, Codex 85%. Outlier: Zed at 38% top-2 box on 5% usage.
The team frames this as continuity, not surprise: the Dart MCP server shipped in mid-2025, official Agent Skills for Flutter and Dart launched at Google I/O alongside agentic hot reload, and a Flutter Antigravity plugin has since shipped. The stated goal: Flutter as a first-party, well-supported target for every major agentic coding tool.
Trust in Flutter versus trust in Google
The new trust section exists to set a baseline before the community model shifts further — Canonical now leads the Flutter desktop roadmap, and a formal contributor ladder has been proposed. Awareness was never the issue: 98% correctly identified Google as Flutter’s creator. But when asked separately, trust in Flutter (83% top-2 box) beat trust in Google (62%) by 20+ points, consistently across every company-size segment.
What builds that trust? Framework performance and stability topped the list at 26%, ahead of community size and activity (20%) and documentation quality (19%). And for a brand-new feature, more respondents said they’d trust it because it’s “battle-tested” by the community (41%) than because Google engineers built it (26%). The team’s conclusion: the credibility is earned independently of the parent company — exactly the kind of trust it wants to protect while opening the maintainer base.
Where responsiveness lags
Agreement that Flutter is proactive on technical issues and responsive to feedback sits at 71% — a full 10+ points below every other reliability-and-safety statement. The structural answer: expanding the Consultancy Program, the GDE network, the Customer Advisory Board, and the Partners Advisory Board as formal feedback channels, while keeping GitHub and Discord primary.
The contributor side has the same shape. Over 70% of currently open PRs come from contributors outside Google, and the only path to trusted roles was informal admission to the flutter-hackers group. Shortly after the survey closed, the team published a four-tier Contributor Ladder (Contributor → Reviewer → Committer → Maintainer) with public, data-driven promotion in a new flutter/contributors repository, replacing the informal sponsorship model and adding a path to project leadership. Because it went public after the survey closed, the team explicitly doesn’t expect it to move these numbers until next cycle.
The four dissatisfaction themes
Asked what’s actually driving dissatisfaction, four themes emerged, and the team’s responses map to concrete releases:
- Platform and ecosystem maturity (44%) — the largest theme, centered on upgrade pain, especially Android’s Flutter/Dart/Gradle/Kotlin/JVM version matrix. The trail of fixes: Flutter 3.29 removed legacy script-based Gradle plugin applications, 3.32 rewrote the Gradle plugin in Kotlin, 3.38 published the official version matrix, 3.41 added AGP 9 support, and 3.44 shipped migrations to built-in Kotlin. On iOS/macOS, 3.44 made Swift Package Manager the default, with 92 of the top 100 iOS plugins migrated.
- Tooling and IDE experience (33%) — mostly demands for deeper Android Studio integration. The team’s AI investment is agent-agnostic (MCP server, Agent Skills), and it acknowledges Android Studio lacked MCP support at Dart MCP’s launch, with work underway.
- Bugs and stability (24%) — pointed at debug tooling: breakpoints persisting after removal, a confusing performance timeline view. The honest answer: no shipped fix; the team asks for reproductions as GitHub issues.
- UI/UX aesthetics (14%) — the smallest theme by volume but the one most tied to Cupertino’s record drop, which got its own deep dive here on the blog.
Why this survey reads differently
Most framework satisfaction surveys are marketing with error bars. This one publishes a subsystem at 61% and calls it “the steepest decline anywhere in the survey,” admits an 11-point novice/expert gap “onboarding tutorials alone cannot solve,” and states plainly that a responsiveness score 10 points below everything else needs a structural fix. Trust, the survey’s own subject, appears to be built exactly the way respondents said it is: through performance, stability, and showing the numbers — including the bad ones.