List your company

Rating on Held-out games (test)

0.49MAEBG-Critic
0.45760.67630.8951.1137Jun 1, 2026

Evaluation results

MethodLinks
BG-CriticStage 2 multi-task training=Excluded2026.060.490.160.368
BG-CriticStage 2 multi-task training=Included2026.060.50.180.36
Qwen3.5-27BBackbone Status=untuned2026.060.760.720.186
Qwen3.5-397B2026.061.051.040.232
Gemini-3.1-Flash2026.061.111.090.25
GPT-5.42026.061.31.290.336