DeepSeek’s V4.1 Flash model drove the near-certain market consensus by processing roughly 41 trillion tokens in the relevant OpenRouter week—more than triple the next closest labs—thanks to aggressive pricing, token efficiency in agentic and coding workloads, and open-weight distribution that accelerates developer adoption. Chinese labs overall captured the majority of routed tokens through sustained gains in cost-performance, pushing U.S. providers like OpenAI and Anthropic into single-digit shares on the platform. A sudden price cut or capability leap from a Western lab, regulatory restrictions on Chinese models, or a shift toward request-based rather than token-based metrics could still alter the outcome before final resolution.
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