Best AI model for Games & Simulations
DeepSeek V4.1 Flash (Venice) ranks first for Games & Simulations with a score of 51.0, passing 50% of its 8 runs.
Lowest average cost per success: Gemma 4 31B Instruct (Venice) at $0.0042.
Based on 33 logged runs across 4 use cases; 7 models had enough runs to rank.
| # | Model | Score | Runs | Pass rate | Cost / success | Avg time | Step-ins / run |
|---|---|---|---|---|---|---|---|
| 1 | DeepSeek V4.1 Flash (Venice) | 51.0 | 8 | 50% (4/8) | $0.03 | 2m | 0.00 |
| 2 | Qwen3 Coder 480B Turbo (Venice) | 95.7 | 4 | 100% (4/4) | $0.0060 | 29s | 0.00 |
| 3 | Grok 4.3 (Venice) | 92.9 | 4 | 100% (4/4) | $0.01 | 25s | 0.00 |
| 4 | Gemma 4 31B Instruct (Venice) | 91.8 | 4 | 100% (4/4) | $0.0042 | 4m | 0.00 |
| 5 | Claude Sonnet 5.5 | 83.6 | 4 | 100% (4/4) | $0.15 | 54s | 0.00 |
| 6 | Claude Haiku 5.5 | 83.6 | 4 | 100% (4/4) | $0.03 | 5m | 0.00 |
| 7 | Claude Opus 5.5 | 75.7 | 4 | 100% (4/4) | $0.90 | 4m | 0.00 |
Only models with 3+ runs here are ranked; all settings of a model are grouped. Scores use the default weights (methodology).
Browse games & simulations use cases → · Filtered leaderboard →