DIY CrusaderBench ranks Gemini 4 Argon #1 at Stronghold strategy, beating Opus 5.5
Hesamation · x · 2026-10-14
- The author built CrusaderBench, testing how well LLMs grow an economy in Stronghold Crusader — a proxy for strategic reasoning, planning, computer use and tool use. Gemini 4 Argon scored 61.7, ranking #1 ahead of Claude Opus 5.5 (51.4); Argon itself called it one of the best-designed game benchmarks it had seen.
- Model notes: Opus is a strong strategic player; Sonnet and Haiku were surprisingly lazy, going idle too early; GPT models consistently solid (GPT-6.1 Sol 46.6, GPT-6 Luna 36.8); Kimi K3 used buildings well but with poor strategy; GLM 5.3 Flash overperformed (27.6); Muse Wonder 3 scored 35.6; Claude Haiku 5.5 last at 20.4.
- Scoring: 80% net worth, 10% population, 10% popularity; 0 = doing nothing, 100 = human player level. Some models still need reruns to filter noise.
Related event: Gemini 4 Argon Tops Homemade CrusaderBench Strategy Benchmark(2 posts)→
More from Models
- Gemini 4 Argon shows strong cyber capabilities: 87% ExploitBench, 68% CWE-Bench — scaling01 · 2026-10-14
- Early users wowed by Gemini 4 Argon's 3D design quality — prajdabre · 2026-10-14
- EmbeddingGemma2 on CPU matches OpenAI's embedding-3-large in 96-config retrieval benchmark — tomaarsen · 2026-10-14
- Physical Intelligence launches Pi API beta for fine-tuning π robotic foundation models — chelseabfinn · 2026-10-14
- Researcher says he had Argon attempt the Indus Valley script, with limited progress — adityakusupati · 2026-10-14
- Leaked Gemini 4 Argon benchmark shows trends matching Astra — scaling01 · 2026-10-14