DeepSeek Releases V4 AI Model with Reduced Costs

Chinese AI company DeepSeek released its long-awaited V4 flagship model featuring drastically reduced costs and extra-long text processing capabilities, though initial benchmarks suggest mixed results. According to Artificial Analysis's Friday report, DeepSeek V4 Pro ranked second among leading open-source models globally, trailing Beijing-based Moonshot AI's Kimi K2.6. While the V4 represents a clear improvement over DeepSeek's predecessor models, analysts indicate it falls short of both domestic Chinese competitors and US rivals, struggling to replicate the market-shaking success of the company's earlier R1 release.

V4 Model Performance Compared to Competitors

• ▪DeepSeek V4 Pro ranked second among the world's leading open-source models according to Artificial Analysis
• ▪DeepSeek V4 Pro marked a clear improvement on its predecessor models
• ▪Beijing-based Moonshot AI's Kimi K2.6 ranked ahead of DeepSeek V4 Pro among open-source models
• ▪Artificial Analysis released a benchmark report on DeepSeek V4 Pro on Friday

DeepSeek's Struggle to Match R1's Success

• ▪DeepSeek is a Chinese artificial intelligence firm
• ▪DeepSeek V4 is described as a long-awaited flagship model
• ▪DeepSeek's V4 model fell short of its domestic and US rivals according to a new analysis
• ▪DeepSeek struggles to replicate the market-shaking success of its earlier R1 release

Perspective of DeepSeek

• ▪DeepSeek V4 has the capability to process extra-long texts
• ▪DeepSeek V4 offers drastically reduced costs compared to previous models

Perspective of AI industry analysts

• ▪DeepSeek V4 fails to replicate the market-shaking impact of the earlier DeepSeek R1 release
• ▪DeepSeek V4 Pro performs below the level of leading US AI rivals according to benchmarks