Fireworks AI previews Ultrafast and adds multi-day prompt caching and background mode for agents

發生什麼事

@FireworksAI_HQ says the company will expand its European capacity with NVIDIA Vera Rubin and Grace Blackwell platforms, with the work starting in 2027. It also introduced Ultrafast in preview, with output at 1,500+ tokens per second. Agent features include prompt caching that lasts several days and a background mode; rate limits were raised too.