POPA ReconDeepSeek V4 Drops With Massive Efficiency Leap, Underpricing GPT-5.5 by Up to 98%
DeepSeek V4 Drops With Massive Efficiency Leap, Underpricing GPT-5.5 by Up to 98% DeepSeek has released V4-Pro and V4-Flash just hours after OpenAI launched GPT-5.5, intensifying the global AI race. Both models are open-weight, support up to 1M token context, and are designed for extreme efficiency at scale, with pricing dramatically undercutting Western competitors. V4-Pro is the flagship model with 1.6T total parameters (49B active per inference), while V4-Flash is a lighter, faster version with 284B total / 13B active. Both use an improved Mixture-of-Experts architecture and new attentio…
The excerpt is quoted from a public Telegram channel, which owns its text.