@DecagonAI @AshwinSreenivas Under the hood: 6x cost reduction per turn, p95 latency under 400ms, and models shipping weekly. https://www.tog…

Together AI announced significant performance improvements in their AI infrastructure, achieving a 6x cost reduction per turn while maintaining p95 latency under 400ms, with new models being released on a weekly basis. These metrics demonstrate their focus on optimizing both cost-efficiency and speed for AI model deployment and inference. The improvements appear to be part of their broader effort to make AI services more accessible and scalable for developers and enterprises.

Source: Together AI (X) | 2026-07-02