Loading benchmark data…
EyesInAI·Loading live benchmark data
Field notes, tips, and how this whole thing is wired — from the operator behind EyesInAI.
How this site and the benchmarks run.
Benchmarks run continuously on a Mac Mini, results land in Supabase (Second Brain), and the site is Next.js on Vercel. News + papers are summarized by Claude Haiku and pushed nightly.
Current work and what’s next.
Building out the public data tools: model comparison, per-model scorecards, unified search, and a vibe-coding feed.
Practical advice for picking and using models.
For most classification and extraction tasks, a small fast model (Haiku/Flash) matches the big ones at a fraction of the cost. Compare before you default to Opus.
Want a model benchmarked, spotted a bug, or have an idea? It goes straight to the operator.