APCA Contrast Explained: What Designers Need to Know Before WCAG 3

If you have ever shipped a color pair that passed every checker and still looked washed out on screen, you have already met the limits of the WCAG 2 contrast ratio. APCA contrast is the proposed fix: a perceptual contrast model built for modern screens that scores readability based on font size, weight, and polarity instead of a single blanket ratio. In this guide, I will break down how APCA contrast works, where it disagrees with WCAG 2 (with real, verified numbers), where WCAG 3 actually stands right now, and how to prepare your color workflow without putting your compliance at risk.

Why the WCAG 2 Contrast Ratio Falls Short

Low-contrast text is still the single most common accessibility failure on the web. The WebAIM Million 2026 report found WCAG 2 contrast failures on 83.9% of the top one million home pages, up from 79.1% the year before, with an average of 34 low-contrast instances per page. After nearly two decades of guidance, the needle is moving in the wrong direction, and part of the reason is the math itself.

The WCAG 2 contrast ratio (success criterion 1.4.3 Contrast Minimum) compares the relative luminance of two colors and outputs a ratio from 1:1 to 21:1. Hit 4.5:1 for normal text or 3:1 for large text and you pass Level AA. It is simple, automatable, and legally entrenched. It is also blind to three things your eyes are not:

• Font size and weight. A 12px thin gray label and a 32px bold heading in the same colors get the identical score, even though one is far harder to read. Spatial frequency, meaning how small and thin the strokes are, is a primary driver of human contrast perception, and the WCAG 2 formula ignores it almost entirely.
• Polarity. The contrast ratio is symmetric: swapping text and background colors produces the same number. Human vision is not symmetric. Light text on a dark background is perceived differently from dark text on a light background, which is exactly why so many dark themes that pass on paper feel muddy in practice.
• Dark colors. The formula systematically overstates contrast when both colors are dark. The APCA documentation puts it bluntly: a 4.5:1 ratio can be functionally unreadable when one color is near black, which is why the WCAG 2 contrast ratio cannot give reliable guidance for dark mode.

The pass/fail cliff creates absurd edge cases, too. Gray #777777 on white scores 4.48:1 and fails AA, while #767676 scores 4.54:1 and passes, even though no human being can tell those two grays apart. Run both pairs through a contrast checker and you will see how arbitrary that threshold is: one hex digit separates compliant from non-compliant with zero perceptual difference. A binary checkbox was never going to capture how reading actually works.

What Is APCA Contrast and How Does It Work?

APCA stands for Accessible Perceptual Contrast Algorithm, developed by Andrew Somers of Myndex Research as part of the candidate readability work for WCAG 3. Instead of a ratio, APCA contrast outputs a lightness contrast value, written as Lc, on a scale from 0 to roughly ±106. A positive Lc value means dark text on a light background; a negative Lc value means light text on a dark background. That sign is not a quirk, it is the polarity awareness the old contrast ratio never had.

The model is built on perceptual uniformity. An Lc value of 60 represents the same perceived contrast whether your palette is pastel or near-black, and doubling the Lc value roughly doubles the perceived contrast. The algorithm also applies a small contrast boost to very dark color pairs, which is precisely where the WCAG 2 math drifts furthest from human perception. One thing APCA does not try to fix is hue discrimination: it measures lightness contrast for readability, so you should still test palettes with a color blindness simulator when color alone carries meaning.

The APCA Lc Levels, Explained

The second big shift is that the target Lc value depends on what the text is for, and on its size and weight. The APCA Readability Criterion defines simple use-case levels (with optional font lookup tables for finer-grained conformance). Here are the core levels, based on reference fonts like Helvetica or Arial:

Notice what this structure buys you. Readability improves where it matters most, because body text is held to a higher bar than the old 4.5:1. Design flexibility improves at the same time, because a 36px bold headline is not forced to meet the same threshold as 14px paragraph text. That trade is the core philosophy behind perceptual contrast: match the requirement to the actual reading task instead of brute-forcing one number everywhere.

APCA vs WCAG 2: Same Colors, Very Different Verdicts

Abstract arguments are easy to dismiss, so I computed both scores for several real color pairs using the official apca-w3 library. This is where the APCA vs WCAG debate stops being academic. The same gray background, #777777, produces my favorite example: WCAG 2 fails white text on it but passes black text on it, while both your eyes and the Lc value say the opposite.

Read that third row again. A dark-mode pairing that sails past WCAG AAA lands below the APCA body-text minimum, a textbook case of the contrast ratio overstating dark-color contrast. And the first two rows show polarity in action: on the identical gray, the readable direction fails the old math and the muddy direction passes it. If you maintain a palette, checking every combination one pair at a time gets tedious fast; a contrast grid lets you score all foreground and background pairings in one pass and spot the fragile combinations immediately.

Want to verify the numbers yourself or wire perceptual contrast into your own tooling? The reference implementation is a few lines of JavaScript:

// npm i apca-w3 colorparsley
import { APCAcontrast, sRGBtoY } from 'apca-w3';
import { colorParsley } from 'colorparsley';

const lc = APCAcontrast(
  sRGBtoY(colorParsley('#FFFFFF')), // text color
  sRGBtoY(colorParsley('#777777'))  // background color
);

console.log(lc.toFixed(1)); // -76.6 (negative: light text on dark background)

Where WCAG 3 and APCA Actually Stand Right Now

This is the part most articles get wrong, so let me be precise about the current status, because the details matter for anyone making compliance decisions.

WCAG 3 (formally the W3C Accessibility Guidelines) remains a Working Draft. The W3C published its latest WCAG 3 Working Draft update in March 2026, and the published roadmap targets a Candidate Recommendation around Q4 2027, with a final W3C Recommendation no earlier than 2028. Regulatory adoption typically lags a final Recommendation by years, which means WCAG 2.1 and 2.2 Level AA remain the operative legal benchmarks today, including for the U.S. ADA Title II web rule whose first compliance deadline for larger public entities arrived in April 2026.

And here is the nuance on APCA itself: it is a candidate method, not an adopted standard. The visual contrast work was moved out of the WCAG 3 Working Draft in July 2023 for further evaluation, and as accessibility consultant Adrian Roselli documented in his April 2026 status review, the current draft states that the WCAG 3 contrast algorithm is yet to be determined. The research behind APCA is substantive, peer-reviewed, and continues to mature, and it remains the most developed perceptual contrast candidate. But nobody can honestly tell you today that APCA is the WCAG 3 contrast method. Treat it as a powerful readability instrument that is ahead of the standard, not as a replacement for it.

How to Prepare Your Color Workflow Today

Given that timeline, the smart play is a two-layer testing strategy: WCAG 2 as your compliance floor, APCA contrast as your readability ceiling. Here is the workflow I recommend.

• Step 1: Keep WCAG 2 as your non-negotiable floor. Every text pair must still clear 4.5:1 (or 3:1 for large text). This is what auditors, procurement teams, and courts measure against, and it will be for years. Validate pairs in a WCAG contrast checker before anything ships.
• Step 2: Add an APCA pass for readability quality. Score your core text styles against the Lc levels above, aiming for Lc 90 on body text and at least Lc 75 where reading matters. When a pair clears the legal floor but lands low on the Lc scale, especially in dark themes, treat that as a design bug even though it is not a compliance bug.
• Step 3: Audit the whole palette, not just hero pairs. Map every approved foreground and background combination in a palette accessibility matrix so the failing combinations are documented before a teammate reaches for them, and check how your real type sizes and weights hold up with a legibility analyzer.
• Step 4: Rebuild dark mode with perceptual contrast in mind. Dark themes are where the two models disagree most. Avoid pure black backgrounds, prefer dark grays around #121212 to #1E1E1E, and push light text brighter than the contrast ratio alone suggests, since WCAG 2 flatters dark pairs. Generating tints and shades of your brand colors with a tint and shade generator makes it easy to nudge a failing step lighter or darker without abandoning the hue.
• Step 5: Bake both thresholds into your tokens. If you maintain a design system, record each token pairing with its contrast ratio and its Lc value, and build new palettes from accessible primitives using a color-safe palette builder. Future-you inherits a system that already satisfies whatever WCAG 3 finalizes.

Common Mistakes When Adopting APCA

• Replacing WCAG 2 testing entirely. The most dangerous mistake. A pair that passes APCA but fails 4.5:1 is still a legal liability today, no matter how readable it looks. The floor stays until regulations change.
• Ignoring the sign of the Lc value. Lc -75 and Lc 75 represent comparable contrast magnitudes in opposite polarities. Tooling that drops the negative sign, or designers who read it as an error, will misjudge every dark-mode measurement.
• Using one Lc number as a blanket threshold. Recreating the old single-ratio habit defeats the entire model. Lc 60 is fine for a 24px subhead and inadequate for 14px body copy; the use-case levels and font lookup tables are the point, not an optional extra.
• Treating experimental tooling as authoritative. Browser developer tools have shipped APCA only behind experimental flags, and implementations can lag the current algorithm revision. For decisions that matter, verify with the official calculator or the apca-w3 library.
• Forgetting that contrast is not the whole job. A perfect Lc value does not rescue text set over a busy image, a paper-thin font, or a UI where color alone encodes state. Perceptual contrast is one instrument in the readability toolkit, not the toolkit.

FAQ: APCA Contrast

Is APCA officially part of WCAG 3?

Not yet. APCA is the leading candidate method developed within the WCAG 3 effort, but the visual contrast section was moved out of the Working Draft in July 2023 for further evaluation, and the current draft says the contrast algorithm is yet to be determined. WCAG 3 itself is not expected to become a final Recommendation before 2028.

What is a good APCA contrast score for body text?

Aim for an Lc value of 90 or higher for columns of body text, with Lc 75 as the practical minimum where readability matters. Larger or bolder text can go lower: Lc 60 covers non-body content text from 24px regular or 16px bold, and Lc 45 covers large headlines. Read the value as an absolute number; the sign only indicates polarity.

Does passing APCA make my site legally accessible?

No. Laws and procurement standards worldwide reference WCAG 2.x Level AA, so the 4.5:1 and 3:1 contrast ratio thresholds remain the compliance requirement. Using APCA guidance in place of WCAG 2 conformance creates legal risk today. Use both: WCAG 2 for compliance, APCA for genuine readability.

Why is my APCA value negative?

A negative Lc value simply means your text is lighter than its background, as in dark mode. APCA computes the two polarities separately because human vision perceives them differently. For threshold checks, compare the absolute value against the level you need, so Lc -76.6 comfortably clears a 75 minimum.

Can colors pass WCAG 2 but fail APCA, or the reverse?

Yes, in both directions, and dark palettes are the biggest source of disagreement. In my tests, #B0B0B0 on #1E1E1E passes WCAG AAA at 7.69:1 yet scores only Lc 57.7, below the APCA body-text minimum. Meanwhile white on #777777 fails AA at 4.48:1 but reaches Lc 76.6, which clears it. Neither model is a subset of the other.

How do I test APCA contrast today?

Use the official APCA calculator at apcacontrast.com, the open-source apca-w3 JavaScript package shown above, or design-tool plugins that implement the current algorithm revision. Keep your existing WCAG checks running alongside; the goal is a second readability signal, not a swap.

Should I stop using the WCAG 2 contrast ratio?

No. Keep it as your compliance floor for as long as regulations reference WCAG 2.x, which will realistically be the case into the 2030s. What you should stop doing is treating 4.5:1 as proof of readability, especially in dark mode, where the formula demonstrably overstates contrast.

The Bottom Line: Test Both, Ship Readable

The WCAG 2 contrast ratio is the law of the land; APCA contrast is a better description of how people actually read. Those two facts can coexist in one workflow. Keep 4.5:1 as the floor your lawyers care about, hold your body text to Lc 90 because your readers care about that, and pay special attention to dark mode, where the gap between the two models is widest. Designers who internalize perceptual contrast now will not need to scramble when WCAG 3 eventually lands; their palettes will already be ahead of it. Start by scoring your current palette, fix the pairs that only pass on paper, and your interface gets more readable this week, years before any standard requires it.

Related tools & resources on 66Colorful:

• Contrast Checker — test any text and background pair against WCAG ratios with live previews
• Contrast Grid — score every pairing in your palette at once
• Palette Accessibility Matrix — document which combinations are safe to use
• Legibility Analyzer — evaluate how readable your text styles really are
• Color Safe Palette Builder — build palettes from accessible primitives
• Focus Indicator Checker — verify your focus states meet non-text contrast needs

Sources: WebAIM Million 2026 (webaim.org/projects/million); W3C WAI, WCAG 3 Working Draft news, March 2026 (w3.org/WAI/news/2026-03-03/wcag3); W3C WAI, Understanding SC 1.4.3 Contrast Minimum (w3.org); Myndex Research, The Easy Intro to the APCA Contrast Method (git.apcacontrast.com); Adrian Roselli, WCAG3 Contrast as of April 2026 (adrianroselli.com); APCA Contrast Calculator (apcacontrast.com).