[DeepSeek V4.1 Flash Third-Party Test: Didn't Reclaim Domestic Model Top Spot, But Costs Only One-Seventh of Kimi and GLM] Beating AI News, Artificial Analysis has completed independent testing of DeepSeek V4.1 Flash, achieving a comprehensive intelligence index score of 40. This surpasses its own V4 Pro 0813's score of 36 but remains below Kimi K3's 44 and GLM-5.3's 45. According to this third-party ranking, DeepSeek has yet to reclaim the top spot among domestic models. However, the cost difference is staggering. Running one Intelligence Index task with V4.1 Flash costs an average of only $0.27, while Kimi K3 and GLM-5.3 are both around $2, a difference of approximately sevenfold. Output speed reaches 197 Tokens/s, significantly higher than Kimi K3's approximately 36 Tokens/s and GLM-5.3's approximately 58 Tokens/s. Its Agent remains its strongest feature. On AutomationBench-AA, V4.1 Flash scored 69%, matching GPT-6 Astra and surpassing GLM-5.3's 62%; it also achieved 84% in the long-context test AA-LCR. The downside is its verbosity—during AA tests, the average output per task was about 89,000 Tokens, 62% more than V4 Pro, which consumed part of its cost advantage. [Original Link]