New AI Models MiniMax M3.1 Flash and Inception Mercury 3 Launch, Promising Cost Cuts and Performance Gains

Key Intelligence

• ■MiniMax has launched its M3.1 Flash AI model, claiming it is 3x faster and cuts AI costs by 70%, with a 'Space Bunny Stealth Model' variant.
• ■The MiniMax M3.1 Flash reportedly outperforms DeepSeek and Opus 5 on key business benchmarks at a lower cost, but is noted for a high hallucination rate.
• ■Inception also released its flagship diffusion LLM, Mercury 3, which is touted for setting a new Pareto Frontier in speed and intelligence.
• ■Comparisons are emerging between the new models, such as Mercury 3 versus DeepSeek V4 Pro, focusing on clock speed and launch pricing.