RuneBench Skill - Hitpoints
Games · 2026-02-26
RuneBench 30-minute Hitpoints skill task. Agents play RuneScape through the SDK and are scored by the best real-game Hitpoints XP per minute achieved in any 15-second window.
Top models (higher is better)
| Model | Score |
|---|---|
| GPT-6 Astra | 3235 |
| GPT-6.1 Sol | 227 |
| Claude Opus 5.5 | 203 |
| Gemini 4 Argon | 170 |
| GPT-6 Sol | 163 |
| Grok 4.5 | 139 |
| Gemini 3.8 Flash | 130 |
| GPT-5.6 Sol | 126 |
| Claude Sonnet 5.5 | 124 |
| Claude Opus 5 | 114 |
| GPT-5.6 Terra | 102 |
| Claude Fable 5 | 101 |
| Gemini 3.7 Flash | 101 |
| Opus 4.8 | 100 |
| DeepSeek-V4-Pro | 100 |
| Grok 4.6 | 89.0 |
| Kimi K2.7 Code | 86.0 |
| Claude Fable 5.1 | 81.0 |
| Gemini 3.5 Flash-Lite | 76.0 |
| Muse Spark 1.3 | 72.0 |
| DeepSeek V4.1 Flash | 70.0 |
| Grok 4.7 | 70.0 |
| Sonnet 5 | 69.0 |
| GPT-5.5 | 69.0 |
| Muse Spark 1.2 | 67.0 |
| GPT-5.6 Luna | 64.0 |
| Claude Haiku 5.5 | 63.0 |
| DeepSeek-V4-Flash-0731 | 63.0 |
| GPT-6 Luna | 63.0 |
| Qwen3.7-Max (source unspecified) | 60.0 |
| Mistral Large 4 Preview | 59.0 |
| Gemini 3.6 Flash | 58.0 |
| Kimi K2.6 | 48.0 |
| Gemini 3.5 Flash | 40.0 |
| GLM-5.3 | 40.0 |
| Qwen3.8 27B | 39.0 |
| Grok 4.3 | 38.0 |
| Muse Spark 1.1 | 38.0 |
| GLM-5.2 | 37.0 |
| Opus 4.6 | 36.0 |