Werewolf Arena / AI town-style game benchmarks — Balroghttps://balrogai.com/

Your landing page’s weakest score is CTA.

How Werewolf Arena / AI town-style game benchmarks — Balrog ranks

  1. 1.Olam Labs66
  2. 2.LMArena (Chatbot Arena)66
  3. 3.Werewolf Arena / AI town-style game benchmarks — Balrog64
  4. 4.TextArena40
Fix CTA
64Growth scoreupdated

category scores

CTA55Mobile web58
category scoresShowHide
CTA55Mobile web58

CTA

55Fair

Copy, size, verb, cost-feel — does the primary CTA earn the click?

Werewolf Arena / AI town-style game benchmarks — Balrog
LMArena (Chatbot Arena)in-product action
Olam Labs

Problems

  • no visual hierarchy among 4 buttons
  • 'Submit' object unclear — submit what?
  • styled same as a dark-mode toggle
  • 'Paper'/'Code' are labels, not actions

Mobile web

58Fair

The phone experience as a first-class citizen.

Mobile web annotations on the mobile capture1234
  1. Author list block — Nine wrapped lines of names push all CTAs and the leaderboard far down the fold.
  2. Four stacked buttons — Paper, Code, Submit, and Mode share equal weight with no visible primary action.
  3. Clipped table header — Leaderboard column header truncates at the fold edge, signaling a table that will not fit the phone width.
  4. Mode toggle button — A theme switch styled identically to conversion buttons wastes a full-height CTA slot.

Value props

68Fair

Title, subtitle, hero — do they sell the product's true value?

Werewolf Arena / AI town-style game benchmarks — Balrog’s headlineBALROGBenchmarking Agentic LLM/VLM Reasoning On Games

Working

Could improve

Trust

68Fair

Proof that the product is real, competent, and safe to try.

Werewolf Arena / AI town-style game benchmarks — Balrog desktop capture for Trust

How Werewolf Arena / AI town-style game benchmarks — Balrog compares

Trust read

Information hierarchy

72Fair

Does the design help or inhibit what the page is trying to say?

Werewolf Arena / AI town-style game benchmarks — Balrog desktop capture for Information hierarchy

What works

What to improve