List your company

Safety on Safety ID

99.81AccuracyGSPO+AIR
59.40669.895580.38590.8745May 20, 2026

Evaluation results

MethodLinks
GSPO+AIR2026.0599.8198.08
GRPO+AIR2026.0598.4684.62
GSPO2026.0597.0270.19
GRPO2026.0596.9271.15
GSPO+V-REx2026.0585.833.2
GPT-5-chat-latest2026.0585.2826.92
GPT-5.2-chat-latest2026.0584.2318.27
GRPO+V-REx2026.0582.1535.4
Llama-4-Maverick2026.0575.385.77
Gemini-3-Pro-Preview2026.0573.176.73
Gemini-2.5-Pro2026.0568.652.88
Qwen-3-14B2026.0565.190
Qwen-2.5-14B-Instruct2026.0560.960