@metaend.eth

Neuralwatt is now available in provider selection for Kimi K3, Qwen 3.6 35B, and DeepSeek V4 Flash 0731, including thinking variants where applicable. Neuralwatt cached-input usage now costs 10% of its regular input rate on supported models, except DeepSeek V4 Flash. Neuralwatt has retired its Kimi K2.6 and Qwen 3.5 routes, so those models are no longer offered through Neuralwatt provider selection; both models remain available through other providers. https://nano-gpt.com/subscription/kUGJhgF9

0 reply

0 recast

2 reactions