Frontier Model Pricing War and Cost Optimization

Aggressive pricing cuts and cost-efficiency claims: OpenAI slashes GPT-5.6 Luna by 80%, Microsoft MAI achieves 50-90% GPU savings, Ramp LLM Router reduces costs 30%, Cursor Agent Swarm 22x reduction. New open-weight models (Kimi K3, Meituan LongCat-2.0) undercut proprietary alternatives.