DEEPSEEK STRIKES AGAIN — V4 MODEL MATCHES CLAUDE AND GPT AT ONE-SIXTH THE COST DeepSeek is playing the same game that sent Silicon Valley into a cold sweat last year: near-frontier AI intelligence at a fraction of the cost, with DeepSeek-V4 now going toe-to-toe with Claude Opus 4.7 and GPT-5.5 at just one-sixth the price. Every AI lab that used billion-dollar training runs as a moat is now being asked, again and loudly, why their model costs six times as much to run. The AI discount aisle is getting more crowded, and the VCs who funded the expensive stuff are smiling a little less broadly. Keywords: DeepSeek V4, AI model cost, open source AI, DeepSeek benchmark, frontier AI