Anthropic’s recent model releases and agent-specific tooling have driven the 88% market-implied probability for the company holding the best AI agent by end of December. Claude Opus 5.5 and the November 9 launches of Fable 5.2 and Mythos 5.2 lead multiple agentic benchmarks in coding, tool use, and long-running workflows, while self-hosted Managed Agents and single-agent architecture improvements have boosted enterprise adoption and reliability metrics. These developments, alongside Anthropic’s reported self-improvement progress and open standards contributions, have widened the gap versus OpenAI’s GPT-6.1 Terra and Google’s Gemini 4 models, which trail in current agent harness evaluations. Traders appear to view Anthropic’s momentum through December as difficult to overcome absent major counter-releases.
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