CraftBench-UE (C++, Unreal MCP): leaderboard

Metric: Task completion (%; C++ tasks completed within 40 minutes over 33 tasks, Unreal MCP editor tools, deterministic build and runtime checks, Claude Code driver). Source: arxiv.org. Saturation forecast: Around December 2026. 7 models tracked.

Top models

#ModelScore
1GPT-5.6 Sol90.91
2Claude Sonnet 590.91
3Grok 4.690.91
4GPT-5.6 Luna87.88
5Gemini 3.7 Flash87.88
6DeepSeek V4 Pro (0813)84.85
7GLM-5.3 Flash69.7

Interactive version: theaggregate.ai/benchmark?slug=craftbench-ue-c-plus-plus-unreal-mcp · How It Works · Data refreshed daily, snapshot 2026-09-26.