Olam LabsYC S26 landing page report.
← Back to YC S26

Score over time

807060504030LMArena (Chatbot Arena) 62 on Jun 30 (Wayback)LMArena (Chatbot Arena) 63 on Jul 15 (Wayback)LMArena (Chatbot Arena) 62 on Jul 24 (Wayback)LMArena (Chatbot Arena) 63 on Aug 6 (Wayback)LMArena (Chatbot Arena) 66 on Aug 24 (Wayback)LMArena (Chatbot Arena) 63 on Sep 5 (Wayback)LMArena (Chatbot Arena) 66 on Sep 20 (Live)TextArena 40 on Apr 17 (Wayback)TextArena 39 on May 20 (Wayback)TextArena 39 on Jun 16 (Wayback)TextArena 40 on Sep 20 (Live)Werewolf Arena / AI town-style game benchmarks — Balrog 64 on Apr 3 (Wayback)Werewolf Arena / AI town-style game benchmarks — Balrog 61 on Apr 14 (Wayback)Werewolf Arena / AI town-style game benchmarks — Balrog 61 on Aug 4 (Wayback)Werewolf Arena / AI town-style game benchmarks — Balrog 64 on Sep 20 (Live)Olam Labs 70 on Jun 19 (Wayback)Olam Labs 66 on Sep 20 (Live)Jun 19Sep 20
Past reportsShow past reportsHide past reports
  1. Sep 20, 202666LiveOverall −4 · Trust −24 · Information hierarchy +4Viewing
  2. Jun 19, 202670First tracked report — strongest CTA 76, weakest Value props 62View →

Trust

42Poor

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

Olam Labs
Werewolf Arena / AI town-style game benchmarks — Balrog
LMArena (Chatbot Arena)
Trust read
Olam Labs"Each opponent is a real AI agent" feature caption on the mobile game card
LMArena (Chatbot Arena)Live "Ask anything…" prompt box with Battle/Auto mode controls showing real product UI
TextArenaNeurIPS 2025 - The MindGame Challenge banner with $500 Modal Labs GPU Credits prize
Werewolf Arena / AI town-style game benchmarks — BalrogLeaderboard table with named models (GPT-6-Astra-Max, Gemini-3-Pro), per-game scores with error bars, and dates

Value props

62Fair

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

Olam Labs’s headline undersellsMulti-Agent ArenaMeasuring how AI agents perform and behave in competitive, social environments

Working

  • Honest, no-hype claim of true value
  • "competitive, social" narrows the category
  • Clear single CTA: Enter Arena

Could improve

  • Zero concrete numbers or proof points
  • Hero image decorates, doesn't demonstrate
  • Buries "play against frontier AI" hook

Mobile web

72Fair

The phone experience as a first-class citizen.

Mobile web annotations on the mobile capture12
  1. Uncollapsed desktop nav — Full four-link desktop nav row is crammed onto the phone, with About nearly clipping the right edge.
  2. Enter Arena button — Primary CTA keeps desktop proportions at ~half column width instead of a thumb-friendly near-full-bleed button.

Information hierarchy

76Fair

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

Information hierarchy annotations on the desktop capture123
  1. Enter Arena button — Semi-transparent dark fill blends into the night-sky hero; the primary CTA reads as a ghost button, not the page's one action.
  2. Headline over clouds — "Multi-Agent Arena" sits directly on busy cloud art with no scrim, so the final letters lose contrast against the bright clouds.
  3. Nav link row — Four equal-weight nav links (Arena, Evaluations, Research, About) are the only other interactive elements and pull attention with no styled conversion anchor in the header.

CTA

78Fair

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

Olam Labs
LMArena (Chatbot Arena)in-product action

Problems

  • low contrast against hero art
  • no payoff hint beyond 'Arena'