Atlas

Benchmarks

← All benchmarks

UnityBench-Mini3D (Easy)

Code · 2026-10-26

The 25 Easy-tier UnityBench-Mini3D tasks. A coding agent modifies a complete initial Unity project offline within one hour, and held-out human-authored EditMode and PlayMode milestone tests run in the Unity Test Runner on the final project. The score is the mean over these tasks of the share of each task's milestone tests that pass, in percent.

Top models (higher is better)

ModelScore
Grok 4.694.6
Claude Opus 594.5
Grok 4.591.5
GPT-5.591.2
GPT-5.6 Terra90.6
GPT-5.6 Sol90.3
Kimi K389.8
GPT-5.6 Luna88.5
GLM-5.284.1
Sonnet 579.6
Kimi K2.7 Code62.8
Gemini 3.7 Flash60.7
Gemini 3.1 Pro Preview43.0
Loading Atlas data…