Open-Source vs Proprietary LLM Competition

DeepSeek-V4 Flash beats GPT-5.6 Luna on DeepSWE at 6x lower cost. Specialized open models beat GPT-5.6 Sol on retrieval at 100x lower cost. AMD Silo AI released Poro 2 with 131K context. Meta's Muse Spark 1.2 updates coding agents but lags behind DeepSeek. The battle for cost-effective, purpose-built models intensifies.