RTSGameBench - Fixed-Field Skirmish Team: leaderboard
RTSGameBench (Kim et al., 2026) collaboration mini-game in Beyond All Reason: the fixed-force skirmish played in teams with no communication channel between allies, so coordination must be inferred from movement, 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 | Claude Sonnet 4.5 | 73 | #244 |
| 2 | GPT-5.2 | 63 | #172 |
| 3 | Kimi K2.5 | 57 | #252 |
| 4 | Mistral Large 3 | 57 | #819 |
| 5 | Gemini 3 Flash | 50 | #159 |
| 6 | Qwen 3 VL 235B A22B (Thinking) | 50 | #451 |
| 7 | Grok 4.1 Fast | 50 | #426 |
| 8 | Qwen 3.5 397B A17B | 37 | #242 |
| 9 | GPT-5 Mini | 30 | #349 |
| 10 | Qwen 3 VL 235B A22B Instruct | 30 | #521 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=rtsgamebench-fixed-field-skirmish-team · How It Works · Data refreshed daily, snapshot 2026-11-12.