Expired50% confidenceBenchmarkExact time
Trial operation data from the Lingxi outfit large model jointly released by Anta Group and Volcano Engine shows an online conversion rate increase of more than 10% and a return rate decrease of more than 10%.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/20/2026
First Seen
Valid until: 10/18/2026(expired)
Sources
Related Claims
Unverified研究表明当ReAct循环超过10步时，模型的任务完成率会出现显著下降60% similarUnverifiedThe power-law exponent between test loss and compute is approximately 0.05, meaning that every 10x increase in compute yields approximately a 12% improvement in model performance.60% similarUnverified在DeepSeek V4线上生产环境实测，推理响应速度提升60%至85%，引擎吞吐量提升51%，高并发场景下吞吐量提升661%59% similarUnverified效率型模型以远低于旗舰模型的推理成本（通常低5-10倍）提供接近旗舰模型80-90%的能力表现57% similarUnverifiedScale AI的研究团队发现，部分模型在被污染的基准上的表现比在新编写的同等难度题目上高出10-15个百分点56% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/570057API
curl https://kongchang.com/api/v1/knowledge/claims/570057MCP
get_claim(id=570057)