List your company

Safe Code Generation on C++ Safe Code Generation Dataset

97.18Safe Code RateClaude Sonnet 4.6
-1.692823.976149.64575.3139Jun 15, 2026

Evaluation results

MethodLinks
Claude Sonnet 4.6Prompting Condition=+Comp. I (SPARK Component I)2026.0697.18
Claude Haiku 4.5Prompting Condition=+Comp. I (SPARK Component I)2026.0695.77
Claude Sonnet 4.6Prompting Condition=Baseline2026.0694.37
GPT-5.4Prompting Condition=+Comp. I (SPARK Component I)2026.0694.37
Claude Haiku 4.5Prompting Condition=Baseline2026.0693.66
GPT-5.4-miniPrompting Condition=Baseline2026.0690.14
GPT-5.4-miniPrompting Condition=+Comp. I (SPARK Component I)2026.0689.44
GPT-5.4Prompting Condition=Baseline2026.0689.44
DeepSeek V4 ProPrompting Condition=+Comp. I (SPARK Component I)2026.0679.58
GPT-5.5Prompting Condition=+Comp. I (SPARK Component I)2026.0666.2
GPT-5.5Prompting Condition=Baseline2026.0664.79
DeepSeek V4 FlashPrompting Condition=+Comp. I (SPARK Component I)2026.0659.86
DeepSeek V4 ProPrompting Condition=Baseline2026.0659.15
DeepSeek V4 FlashPrompting Condition=Baseline2026.0621.13
GPT-5.4Prompting Condition=Plain2026.067.04
Claude Haiku 4.5Prompting Condition=Plain2026.065.63
Claude Sonnet 4.6Prompting Condition=Plain2026.064.93
DeepSeek V4 ProPrompting Condition=Plain2026.064.93
GPT-5.5Prompting Condition=Plain2026.064.23
DeepSeek V4 FlashPrompting Condition=Plain2026.063.52
GPT-5.4-miniPrompting Condition=Plain2026.062.11