Four Chinese AI Labs Drop Frontier-Capable Open-Weights Models in 12 Days

GLM-5.1, MiniMax M2.7, Kimi K2.6, and DeepSeek V4 all launched within a 12-day window in early May — each matching Western frontier performance on agentic coding benchmarks at less than a third of Claude Opus 4.7's inference cost. Inference cost is collapsing faster than capability is growing, and Chinese labs are leading that compression.

Source: AI Tools Recap