Benchmarks / Agentic
EBR-bench (Earthborne Rangers, Epoch AI)
Epoch AI's agentic game benchmark built on the Earthborne Rangers card game: long-horizon planning across a campaign. Score is the share of the scenario completed. Data: Epoch AI, CC BY.
EBR-bench (Earthborne Rangers, Epoch AI)
Higher is better · best effort per model
full table below
| # | Model | Effort | Score |
|---|---|---|---|
| 1 | GPT-6 Astra | max | 76.2% |
| 2 | Opus 5.5 | max | 71.4% |
| 3 | Gemini 4 | high | 64.8% |
| 4 | Fable 5.1 | max | 57.1% |
| 5 | GPT-6.1 Sol | max | 54.3% |
| 6 | GPT-6 Sol | max | 53.3% |
| 7 | Opus 5 | max | 45.7% |
| 8 | GPT-5.6 Sol | max | 44.8% |
| 9 | Haiku 5.5 | max | 43.8% |
| 10 | Fable 5 | max | 39.5% |
| 11 | GPT-5.5 | xhigh | 34.3% |
| 12 | Grok 4.6 | xhigh | 30.5% |
| 13 | Opus 4.8 | max | 28.6% |
| 14 | GPT-5.4 | xhigh | 25.4% |
| 15 | GPT-5.2 | xhigh | 23.0% |
| 16 | Opus 4.7 | max | 19.0% |
| 17 | Opus 4.5 | high | 14.3% |
| 18 | Gemini 3.1 Pro | high | 14.3% |
| 19 | Opus 4.6 | max | 12.7% |
| 20 | GPT-5 | high | 12.7% |
| 21 | GLM 5.2 | max | 9.5% |
| 22 | Qwen 3.7 Max | max | 9.5% |
| 23 | Opus 4.1 | high | 7.9% |
| 24 | Gemini 3.5 Flash | high | 4.8% |
| 25 | Kimi K2.6 | max | 2.4% |
| 26 | Sonnet 4.5 | high | 2.4% |