CrusaderBench Tests 15 Models on Stronghold Crusader, GPT-6.2 Sol Leads
CrusaderBench evaluates AI agents by playing Stronghold Crusader via screenshots and mouse-keyboard control; GPT-6.2 Sol tops the leaderboard with 79 points among 15 models.
2026-11-20 ~ 2026-11-20 · 2 related posts
- CrusaderBench: GPT 6.2 Sol Tops Stronghold Benchmark at 79, Gemini 4 Neon Close — Hesamation · 2026-11-20
- CrusaderBench Makes AI Agents Play Stronghold Crusader; GPT-6.2 Sol Tops Board — Hesamation · 2026-11-20