Every category where a model named Split for a mid-market B2B company. Share is first choices across the direct, paraphrase, budget and scale prompts; rank is within every product named in that category.

Category	Function	Share	Rank	Negative rate	Labels	Quadrant	Since September 2026
Feature flags	Developer platform	0%	9 of 35	18%	22	accepted challenger	=heldSince September 2026: 0% → 0%, ±0 points. Inside the 11-point floor: within noise. Read over the models both editions asked.

Movement

Since September 2026, Split held within the floor in its one category: no change cleared 11 points.

Best category · Feature flags · by segment

Small business▼−2Since September 2026: 2% → 0%, −2 points. Inside the 11-point floor: within noise. Read over the models both editions asked.

0%from 2% in September 2026

Inside the floor by 9 pointsRank 9 → 29 of 30

Mid-market=heldSince September 2026: 0% → 0%, ±0 points. Inside the 11-point floor: within noise. Read over the models both editions asked.

0%from 0% in September 2026

Inside the floor by 11 pointsRank 10 → 9 of 33

Enterprise▼−2Since September 2026: 2% → 0%, −2 points. Inside the 11-point floor: within noise. Read over the models both editions asked.

0%from 2% in September 2026

Inside the floor by 9 pointsRank 8 → 9 of 32

All categories · Mid-market

Category	September 2026	Now	Change	Reading	Rank
Feature flags	0%	0%	=heldSince September 2026: 0% → 0%, ±0 points. Inside the 11-point floor: within noise. Read over the models both editions asked.	Within noise	Rank 10 → 9 of 35

Shares here are read over the models both editions asked, so they can differ by a point or two from the standing above, which counts every model in this edition.

The floor is 11 points of share, measured: how far the models move a leader on their own when the same questions are asked twice with nothing changed. A larger change is movement; a smaller one is noise, and both are shown. Movement is read over the twelve models both editions asked; GPT-6 Luna, Muse Glimmer 30B joined this edition and are in the standing but not yet in the comparison. How the floor is measured · The editions

By model

How each model treated Split across every prompt where it was named for a mid-market B2B company. Fourteen models, six prompts per category.

ShowHide

Model	Alternative	Mention	Negative	Labels
Claude Haiku 4.5	0	1	0	1
GPT-5.4 mini	1	1	0	2
Gemini 3.5 Flash	0	0	1	1
Perplexity Sonar	1	0	0	1
Grok 4.20	2	1	1	4
Mistral Small	0	1	0	1
DeepSeek V4 Flash	1	1	1	3
Llama 4 Maverick	1	0	0	1
Qwen 3.7 Flash	0	1	0	1
Kimi K2	1	2	0	3
GLM 4.7 FlashX	0	0	0	0
MiniMax M2.5	1	1	0	2
GPT-6 Luna	0	0	0	0
Muse Glimmer 30B	1	0	1	2

By framing

Which of the six questions produced the naming. By model says how often; this says asked what. The first-choice count on the right carries the marks of the models that produced it.

FramingLabels by classFirst choices

Direct12 labelsNone

Paraphrase13 labelsNone

Comparative15 labelsNone

Budget-constrained6 labelsNone

Scale-constrained18 labelsNone

Negative14 labelsNone

First choiceAlternativeMentionNegative78 labels in all, every segment counted; 0 of the 0 first choices count toward share, since the comparative and negative framings do not. The bar is one segment per label class, to scale within the framing.

What the models said for it

Verbatim evidence the judge attached to positive labels.

“Choose Split/Harness if experimentation and measuring the business impact of every release is your top priority.” DeepSeek V4 Flash · Feature flags · comparative prompt · alternative

“Split - a feature management tool that offers features such as feature flags, experimentation, and analytics.” Llama 4 Maverick · Feature flags · paraphrase prompt · alternative

“Choose Harness/Split if you want flags plus experimentation/release management in a larger DevOps platform.” GPT-5.4 mini · Feature flags · comparative prompt · alternative

“Best for Experimentation ... Trade-off: Experimentation features gated to Business/Enterprise tiers” Kimi K2 · Feature flags · direct prompt · alternative

And against it

Verbatim evidence attached to negative labels. A warning on a product with few labels is a warning; on a product with many, it is one voice among them.

“during migration periods, users frequently report UI friction, changes to user roles/permissions, and uncertainty regarding the long-term support of standalone features” Gemini 3.5 Flash · Feature flags · negative prompt · soft negative

“Split.io: Solid for experiments, but custom/enterprise pricing; fewer specific complaints, though tied to Harness acquisition.” Grok 4.20 · Feature flags · negative prompt · soft negative

“The acquisition has raised questions about roadmap and pricing alignment with Harness FME.” Muse Glimmer 30B · Feature flags · negative prompt · soft negative

“Pricing opacity pushed many mid-market customers to alternatives” DeepSeek V4 Flash · Feature flags · direct prompt · soft negative

Named alongside

The products named in the same answers as Split, over the 78 answers that named it. Took the first choice instead counts the answers where the other product was the first choice and Split was named but was not.

ShowHide

ProductSame answerTook the first choice insteadHead to head

LaunchDarkly76 of 7828Not among the top eight

Unleash66 of 7811Not among the top eight

Flagsmith56 of 7812Not among the top eight

ConfigCat39 of 7810Not among the top eight

GrowthBook39 of 782Not among the top eight

Statsig37 of 782Not among the top eight

PostHog Feature Flags19 of 780Not among the top eight

Optimizely15 of 780Not among the top eight

DevCycle9 of 780Not among the top eight

Harness9 of 780Not among the top eight

A head-to-head page exists where both products are among a category's top eight. The other rows are the same fact without a page behind them, so they link to the product instead.

What carried it into the answer

The sites and pages cited by the answers that named Split. A fact about retrieval, not a lever on the model.

Citations exist only for the models that return a source list, five of the fourteen in this edition, so these counts come from 63 of the 78 answers that named Split and are not a share of its labels.

Domains cited

flagsmith.com45

growthbook.io42

statsig.com27

getunleash.io25

guideflow.com24

guptadeepak.com24

harness.io24

launchdarkly.com20

configcat.com18

linkedin.com18

267 of the 267 domain citations in answers naming Split came from somebody else's page.

Pages cited

flagsmith.com/blog/top-7-feature-flag-tools26

guptadeepak.com/tools/top-5-feature-flag-management-tools-202624

harness.io/blog/feature-flag-tools-compared-10-best-platforms-for-safer-releases20

growthbook.io/blog/unleash-alternatives18

growthbook.io/insights/best-feature-flag-tools-for-developers18

dupple.com/learn/best-feature-flag-tools16

guideflow.com/blog/feature-flagging-software16

linkedin.com/pulse/best-feature-flag-management-tools-rajesh-kumar-j0doc16

Pages are listed as the models cited them.

Search and answers

split.io is not in Google's top ten for any of the category's searches. In the answers, Split takes 0% of first choices and Flagsmith takes 25%.

ShowHide

In search

Google, US estimates

Position for “feature flag platform”

–

not in the top ten

Position for “feature flag platforms”

–

not in the top ten

Searches for its name, Google: 246,000 a month (“split”)
AI search demand for its name, est.: 342,492 a month
Organic visits to its site: about 1,066 a month
Searches its site ranks for: 1,245 · 33 in the top three
Sites linking to it: 1,757
Paid Google search: No ads found in the estimate; that does not mean it runs none

Ads on Google

None found in Google's ad transparency records for the US

In answers

This edition

Share of first choices

rank 9 of 35 in feature flags

Segment leader

25%

Flagsmith

First choices: 0 across its categories
Named in: 78 answers
Its own site cited: in 63 of the answers that named it

Search figures are US estimates from DataForSEO, read September 28, 2026; AI search demand is its modeled, directional estimate, not a count of queries to any assistant. The answers are this edition's. Two measurements side by side: neither is read as the cause of the other.

Names read as Split

What the judge wrote, as written, with how often. The vendor table decides that these count as Split; a claim can dispute any of them.

Split.io 6Split (now Harness FME) 2Harness Feature Flags / Split 1Split (now Harness Feature Management) 1Split (now part of Harness) 1Split.io (Harness) 1

Follow Split

An email the morning each edition publishes: where this product moved, where it held, and by how much against the noise floor. One address, confirmed by a click; a stop link in every email.

Already following? Everything you follow, with a stop for each.

The company

Split is its own company.

ShowHide

In its own words

Stated by the vendor, not checked

Positioning: Split.io feature flags connects critical impact data and alerts you whether your software changes are making things better or worse. Feature Management & Experimentation is a shift to safer, faster, more impactful software delivery.
split.io
Free plan or trial: Free Account split.io
Customers named: ExperianImperfect FoodsSkillshare
Not stated on the pages read: Who it is for, price, starting price, integrations, certifications, hosting

What Split's own pages state, read September 28, 2026: split.io. A claimed page can correct any of them.

Is this your product?

Claim this page

Claiming is free and changes nothing in the data. A claimed page shows a verified contact who is told when each edition publishes and when Split's standing changes by more than the noise floor; the right to propose corrections to the vendor table, meaning names the judge wrote that should or should not read as Split, applied by version and listed in the change log; and a one-line description supplied by the vendor and marked as such.

What a new claim receivesHide what a new claim receives

A new claim receives the current edition's vendor brief for Split by email, built from the raw record of the edition. It shows:

where Split is named, by buyer and by framing, and which cells hold its first choices;
the claims the models make when they name it, ranked, with the strongest and the weakest quoted;
its vocabulary against the segment leader's, and the pages the models cited;
who was chosen in the answers that did not name Split, and every reason the record gives;
a battlecard for each top rival: the head-to-head split, why they win, and the reservation quoted against them;
one page of published figures cleared to show a buyer.