Anthropic holds a narrow market edge on LiveBench overall by end-November because its Claude Fable 5.2 and related Opus/Sonnet variants currently top multiple public snapshots with scores around 94.5% and strong results across coding, reasoning, and long-horizon agentic tasks. OpenAI’s GPT-6 Astra and GPT-6.1 Sol sit within a few points behind, creating tight implied probabilities near 50% for several frontrunners and leaving room for a late submission or configuration tweak to flip the ranking. Google’s Gemini 4 models trail further back, while Chinese labs and others post lower absolute scores despite competitive efficiency. With only weeks remaining, any fresh Anthropic update or OpenAI benchmark push could shift trader consensus before resolution.
基于Polymarket数据的AI实验性摘要。这不是交易建议,也不影响该市场的结算方式。 · 更新于1H6H1D7D30DALL
View resolved






















警惕外部链接哦。
警惕外部链接哦。
常见问题