Fable 5.2 在 ARC-AGI-2 得分 93.8%
摘要
@AnthropicAI 在 ARC-AGI-2 上,Fable 5.2 的 93.8% 是我们见过的第三高模型得分,仅次于 GPT-6 Astra 的 95.0% 和 GPT-6.1 Sol 的 94.2%。 完整结果:https://arcprize.org/results/anthropic-claude-fable-5-2
摘要由模型根据原文撰写,以原文为准。
正文 · 原文
@AnthropicAI On ARC-AGI-2, Fable 5.2's 93.8% is the third-highest model score we've seen, after GPT-6 Astra's 95.0% and GPT-6.1 Sol's 94.2%.
Full results: https://arcprize.org/results/anthropic-claude-fable-5-2
来源:ARC Prize · x.com
引用原文
arcprize. @AnthropicAI On ARC-AGI-2, Fable 5.2's 93.8% is the third-highest model score we've seen, after GPT-6 Astra's 95.0% and GPT-6.1 Sol's 94.2%.[EB/OL]. (2026-11-10)[访问日期]. https://x.com/arcprize/status/2119871082756800545.