Anthropic’s recent release of Claude Fable 5.2 has driven trader sentiment on the 2026 Humanity’s Last Exam market, with the model posting a leading 74.3% on the main Artificial Analysis HLE leaderboard as of mid-November. This large language model benchmark comprises roughly 2,000 expert-authored questions across reasoning and knowledge domains, designed to resist simple retrieval. Anthropic variants now occupy the top positions on multiple HLE variants, ahead of Moonshot AI’s Kimi K3.1 Pro near 70% on tool-augmented runs and OpenAI’s GPT-6 Astra in the mid-50s. Competitive releases from Google Gemini and open-weight labs like DeepSeek continue to close gaps, while further model updates or refined evaluation protocols before year-end remain key swing factors for the highest reported score.
基於Polymarket數據的AI實驗性摘要。這不是交易建議,也不影響該市場的結算方式。 · 更新於1H6H1D7D30DALL
View resolved

警惕外部連結哦。
警惕外部連結哦。
Frequently Asked Questions