Artificial Analysis 对比六款模型幻觉检查器
We compared six hallucination checkers on the same deliverables from a fixed subset of 20 tasks and eight models. We ran our two-stage check using GPT-6 Sol (hi
Artificial Analysis 在 20 个任务、8 个模型的相同交付物上对比六款幻觉检查器，并选定 GPT-6 Sol (high) 用于生产。GPT-6 Sol 与 GPT-6 Luna 识别出的实质性幻觉更多，Claude Sonnet 5.5 和 Gemini 3.8 Flash 明显更少；总计 Opus 认定 99 个、Grok 219 个、Sonnet 57 个、GPT-6 Sol 470 个。六款检查器均未在 GPT-6 Astra 输出中发现实质性幻觉，但该对比只反映检查器行为差异，计数本身不能证明准确率或排除自我偏好。
正文
当前显示原文，中文全文译文尚未就绪。
We compared six hallucination checkers on the same deliverables from a fixed subset of 20 tasks and eight models. We ran our two-stage check using GPT-6 Sol (high), GPT-6 Luna (high), Grok 4.7 (high), Claude Opus 5.5 (high), Claude Sonnet 5.5 (high) and Gemini 3.8 Flash (high), and selected GPT-6 Sol (high) for production usage.
Across this 20-task subset, GPT-6 Sol and GPT-6 Luna generally identified more material hallucinations, while Claude Sonnet 5.5 and Gemini 3.8 Flash identified far fewer. Claude Opus 5.5 fell between Grok 4.7 and Claude Sonnet 5.5 in every model row. In total, Opus upheld 99 material hallucinations, compared with 219 for Grok, 57 for Sonnet and 470 for GPT-6 Sol. GPT-6 Sol identified more material hallucinations despite the check’s conservative approach to flagging errors, which excludes general legal knowledge not contained in the source documents.
All six checkers found no material hallucinations in GPT-6 Astra’s outputs, while GPT-6 Sol ranged from 0 to 0.20 per task. Both were among the models with the fewest material hallucinations under every checker. This comparison shows differences in checker behavior; the counts alone do not establish accuracy or rule out self-preference.