Models
Every model tested on Crusader Bench, in leaderboard order on v1.5 and v1.6, with its score on each map and over every map. Each links to its own page, with every game it played and the models to compare it with.
| Model | Every map | Oasis by the Sea | Large Barren Desert | The Sinkhole | Cost | Games |
|---|---|---|---|---|---|---|
| Claude Opus 5.5Anthropic | 50.4 | 52.0 | 67.6 | 31.4 | ~$7.54 | 3 |
| GPT-6.1 SolOpenAI | 46.6 | 43.4 | 47.6 | 48.8 | $2.90 | 3 |
| GPT-6 LunaOpenAI | 36.8 | 39.0 | 45.9 | 25.5 | $0.38 | 4 |
| Claude Sonnet 5.5Anthropic | 34.2 | 43.6 | 21.0 | 38.0 | ~$1.89 | 4 |
| GLM 5.3 FlashZ.ai | 27.6 | 33.8 | 33.7 | 15.2 | $0.37 | 6 |
| Kimi K3Moonshot AI | 23.9 | 31.1 | 19.4 | 21.1 | ~$7.30 | 3 |
| Claude Haiku 5.5Anthropic | 20.4 | 26.6 | 26.6 | 8.0 | ~$0.17 | 5 |