Feature Flag Pricing in 2026: What You'll Actually Pay

Feature flag pricing is hard to Google because every vendor meters a different thing. Seats. Client-side MAUs. Service connections. Flag requests. "Contact sales." Two quotes for the same team can differ by 10x and both be "correct" on the pricing page.

I am building Betterflag, so this is not a neutral buyer's guide. It is the teardown I wanted when I was the buyer. Numbers below are published list prices as of mid-2026. Verify before you sign. Side-by-sides live on /vs.

The meters, in one place

| Meter | What grows the bill | Who uses it | |---|---|---| | Seats | Hiring, plus anyone who needs to see a rollout | GrowthBook Cloud Pro ($40/seat), older enterprise SKUs | | Client-side MAUs | Your product succeeding | LaunchDarkly ($8.33–$10 per 1,000 MAUs) | | Service connections | Each SDK instance × environment (pods, workers, serverless) | LaunchDarkly (~$10 per connection / mo) | | Flag requests | Every evaluation that hits their API | PostHog (after 1M free), Flagsmith | | Evaluations (one meter) | Flag decisions served | Betterflag ($9.99 / 1M, $24.99 / 5M, $99.99 / 50M) | | Events | Analytics around the flags | Statsig (flag checks free; events metered) | | Your time | Upgrades, uptime, SDKs | Any self-hosted OSS |

If the meter is not the thing that costs the vendor money, you are paying for a proxy. Evaluations are the thing that costs money. MAUs and seats are easier to sell. I argued this at more length in why LaunchDarkly bills your architecture.

Worked example: 10 people, 25k MAU, ~5M evaluations

Assumptions: a consumer app, 5 backend services, staging + production, 25,000 monthly active users, flags on the main flows. We use ~200 evaluations per MAU per month as a planning number (page loads plus API checks). 25k × 200 = 5M evaluations. If you cache aggressively, you will be lower. If you evaluate on every render, you will be higher.

| Tool | What you pay (list, monthly) | Notes | |---|---|---| | LaunchDarkly Foundation | ~$50 connections + ~$250 MAUs ≈ $300 | 5 services as 5 connections; MAUs dominate. Full vs | | PostHog flags | ~$235 after the free 1M | Local evaluation can cut this. Full vs | | GrowthBook Cloud Pro | $400 | 10 × $40 seats. Self-host is $0 + ops. Full vs | | Flagsmith Start-Up | $45 base, then expensive overage; 3-seat cap | 5M requests is a lot of overage at $7/100k. Full vs | | Statsig | Flag checks free; Pro from $150 if you need team features | Buy it for experiments. Full vs | | Betterflag Launch | $24.99 | 5M evaluations included, unlimited seats |

This is why "how much do feature flags cost" has no one-line answer. The same job is $25 or $400 depending on whether you are buying seats, MAUs, or evaluations.

Scale the app to 100k MAU (~20M evaluations). LaunchDarkly MAU line alone is about $833–$1,000/mo. Betterflag's cheapest bill is Launch plus 15M overage at $84.99/mo, not the $24.99 included price. Scale is $99.99 if you want 50M included. That gap is the commercial argument. Do the evaluation math.

Honest cases where you should not buy Betterflag

You already pay for PostHog analytics and stay under 1M flag requests. Flags are effectively free. Stay. Move later if volume grows.

You run a real experimentation program. Statsig's stats engine is excellent. Flags ride along. Betterflag will not grow a CUPED implementation to win you.

You must self-host. Unleash, Flagsmith, GrowthBook. We are hosted-only. That is a real requirement for some teams, and it is a door we do not open.

You are a 500-engineer org with a platform team and a compliance matrix. LaunchDarkly is built for you. You are not overpaying for a boolean. You are paying for a vendor that looks like the rest of your stack.

How to forecast any vendor's bill

1. Pick the meter. If there are two, budget both at 3× current traffic.
2. Count service connections honestly: services × environments, including preview if they bill those.
3. Convert MAU to evaluations if you are comparing to an evaluation-priced tool. 100–300 evaluations per MAU per month is a sane band for an app that flags main flows. Publish the conversion next to the quote so nobody can accuse you of a trick.
4. Ask what an agent or a CI job costs. If the answer is a seat, the model predates your workflow.
5. Ask where the audit log, approvals, and SCIM live. If the answer is "Enterprise, custom," your kill switch has an upsell.

One meter, on purpose

Betterflag pricing is evaluations only. Unlimited flags, seats, environments on every tier. Starter $9.99/mo (1M), Launch $24.99 (5M), Scale $99.99 (50M), overage published at $5 / $4 / $3 per extra 1M. No MAU math, no connection audit, no seat spreadsheet.

Use the bill calculator to convert your MAU and services into both meters. The conversion (200 evaluations per MAU by default) is on the page so nobody can accuse the quote of a trick.

Alpha waitlist members lock in 50% off for life. That is the deal while we are pre-launch. Compare vendors if you want the long tables. Come back to this post when a pricing page has four asterisks and a "talk to sales" where the number should be.