RTSGameBench - Multi-Front Defense: leaderboard
RTSGameBench (Kim et al., 2026) spatial and temporal reasoning mini-game in Beyond All Reason: defend several objectives against attacks arriving from different directions at staggered times, with fixed forces and no production, played through the RTSGameAgent harness (FSM group management plus agentic memory, the same VLM for memory and decisions). Win rate over 30 runs, in percent; higher is better.
Metric: Win Rate (%; 30 runs). Source: arxiv.org. Saturation forecast: Around February 2027. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Kimi K2.5 | 77 | #252 |
| 2 | Gemini 3 Flash | 60 | #159 |
| 3 | Qwen 3 VL 235B A22B (Thinking) | 53 | #451 |
| 4 | Claude Sonnet 4.5 | 33 | #244 |
| 5 | GPT-5.2 | 30 | #172 |
| 6 | GPT-5 Mini | 23 | #349 |
| 7 | Qwen 3.5 397B A17B | 20 | #242 |
| 8 | Grok 4.1 Fast | 10 | #426 |
| 9 | Qwen 3 VL 235B A22B Instruct | 0 | #521 |
| 10 | Mistral Large 3 | 0 | #819 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=rtsgamebench-multi-front-defense · How It Works · Data refreshed daily, snapshot 2026-11-12.