List your company

Moral Reasoning on Moral ID

85.51AccuracyGRPO+AIR
56.3963.9571.5179.07May 20, 2026

Evaluation results

MethodLinks
GRPO+AIR2026.0585.5159.85
GSPO+AIR2026.0584.5765.69
GPT-5.2-chat-latest2026.0583.2129.2
GSPO2026.0581.0240.15
GPT-5-chat-latest2026.0580.8122.63
GRPO2026.0575.3934.31
Gemini-3-Pro-Preview2026.0567.7810.22
Qwen-3-14B2026.0567.3610.95
Llama-4-Maverick2026.0564.965.11
Qwen-2.5-14B-Instruct2026.0562.153.65
GSPO+V-REx2026.0561.510.8
GRPO+V-REx2026.0558.27.15
Gemini-2.5-Pro2026.0557.517.3