WebKit closed real gaps in Safari 26.5 and the Safari 27 beta — five separate anchor-positioning bugs, several scroll-driven-animation fixes, and five block-in-inline layout fixes all landed in the same release. But a few named gaps are still live in production Safari today: CSS Masonry layout isn't supported at all, and anchor positioning, while much better, still has rough edges the spec doesn't anticipate. For teams maintaining WKWebView screens inside a broader iOS app, this matters even without a Safari tab in sight: anything rendered through WKWebView inherits the same rendering engine, the same fixed bugs, and the same open ones — exactly the kind of platform quirk that ends up on the desk of a native iOS development team once a workaround stops being worth maintaining.
That split — genuinely fixed vs. genuinely still open — is worth being precise about, because "Safari is behind" has become a reflexive complaint that obscures which specific thing actually broke a specific build. This piece names the gaps that are still open, names the ones WebKit just closed, and ends with a short list of what to actually test before you ship.
When one of these gaps is severe enough that a web-based screen genuinely can't deliver the interaction a product needs — not "slightly worse," but structurally blocked — that's a native development decision, not a browser-testing one, and it's worth treating it as such rather than layering workaround on top of workaround inside a WebView.
The Interop project tracks how consistently browsers implement a shared set of web-platform tests each year, and WebKit's score has been closing the gap with Chromium and Gecko rather than standing still — some individual feature areas (CSS Nesting among them, below) jumped dramatically between the 2025 and 2026 scoring rounds. Treat "Safari is behind" as a moving measurement, not a fixed verdict: some gaps that were real six months ago are closed, and a few new or stubborn ones remain open at the same time.
No. CSS Masonry layout is not supported in Safari at all as of late 2026, alongside the other major engines' partial-at-best status. If a layout genuinely needs masonry-style item placement today, the practical options are a manual CSS Grid with explicit row/column placement, or a small JavaScript layout library — there's no native fallback worth waiting on in the current release cycle.
@scope Work Reliably Now?Mostly. CSS @scope has shipped in Safari since 17.4, and for straightforward scoped-style use cases it behaves as the spec describes. The edge cases worth testing are at the scope boundary itself: donor and limit selectors combined with deeply nested slotted content, and scoped rules interacting with cascade layers, are where inconsistent behavior has been reported across engines, Safari included. Don't assume a passing smoke test on a simple component means the boundary logic holds for a more complex one.
Before Safari 26.5, CSS anchor positioning had five specific, reproducible gaps against the spec: media-query re-evaluation didn't re-trigger correctly on viewport resize, anchor chains of three or more elements didn't resolve properly, the anchor() fallback value rejected a unitless zero it should have accepted, elements with display: contents didn't correctly participate in anchor scoping, and a fixed-position element anchored to a sticky-positioned child didn't track it as expected. WebKit's own release notes for Safari 26.5 list exactly these five as fixed, alongside four scroll-driven-animation fixes and five block-in-inline layout fixes landing in the same release — this is the clearest example of the "fixed, not frozen" pattern this article is making the case for.
On top of that, the Safari 27 beta adds further anchor-positioning work: transform-aware anchor positioning (so a transformed ancestor no longer throws off anchor math), the position-anchor: normal and none keywords, and the anchor-valid and anchor-visible pseudo-classes for styling based on whether an anchor reference currently resolves. None of that is stable yet — it's beta — but it's the direction the gap is closing in.
Two items worth separating out, because lumping them in with "Safari's problems" is inaccurate. CSS Mixins and Functions — the proposed way to define reusable, parameterized CSS logic — aren't supported in any browser engine yet, Safari included; this isn't a WebKit-specific shortfall, it's a feature that hasn't shipped anywhere.
Conversely, CSS Nesting's relaxed syntax is a gap that WebKit has already closed: its Interop score for that feature area jumped from 43 to 99 between the 2025 and 2026 scoring rounds. Practically, that means nested selectors using the relaxed syntax (without requiring the & prefix in every position) now behave consistently in current Safari — a case worth citing specifically against the generic "Safari is always behind" narrative, because here it demonstrably isn't.
Two gaps show up specifically on iOS rather than desktop Safari. The Popover API has a known issue on iOS where tapping outside a popover doesn't reliably dismiss it the way it does on desktop — a difference that only surfaces when testing on an actual iOS device, not the desktop Safari simulator most people reach for first. Separately, iOS PWAs installed to the home screen lag behind native apps on Push API delivery and background sync, which affects any installable web app that depends on timely background updates rather than a user-initiated refresh.
@scope rules: test the donor/limit boundary with deeply nested or slotted content, not just a simple flat case.color-scheme declarations: cross-check against what's covered in Safari's Adaptive Dark Mode coverage, since an undeclared color-scheme interacts with more than one WebKit behavior at once.No. CSS Masonry isn't supported in Safari as of late 2026. A manual CSS Grid with explicit placement, or a small JavaScript layout library, are the practical alternatives today.
@scope safe to use in production Safari?Mostly yes — it's been supported since Safari 17.4 for straightforward cases. The edge cases to test are at the scope boundary itself, particularly with nested or slotted content and cascade layers.
WebKit fixed five specific bugs: media-query re-evaluation on resize, anchor chains of three or more elements, the anchor() fallback rejecting unitless zero, display: contents elements in anchor scoping, and fixed-position elements anchored to sticky children.
No. This isn't a Safari-specific gap — no browser engine has shipped support for CSS Mixins and Functions yet.
Yes. WebKit's Interop score for CSS Nesting's relaxed syntax jumped from 43 to 99 between the 2025 and 2026 scoring rounds, meaning nested selectors using that syntax now behave consistently in current Safari.
Not quite — there's a known issue on iOS where tapping outside a popover doesn't reliably dismiss it, unlike desktop Safari. Test this specifically on a real iOS device.
When the gap blocks a core interaction rather than just degrading it cosmetically — and when the screen already lives inside a WKWebView, which inherits the exact same WebKit behavior as Safari itself. At that point, a native rebuild of that specific screen is usually a more durable fix than another layer of workaround code.