ModelsEvery model.
Every model.
Every chip.
Every model gets the best inference performance on every hardware.
DeepSeek V4.1 Flash
DeepSeek · Language modelDeepSeek's open-weight mixture-of-experts language model with a context window of up to 1M tokens.
1M contextOpen weights
MiniMax-H3
MiniMax · Video + audio generationOpen-weight video and audio generation from text, frames or references. One request returns a synchronized MP4.
768p4 to 15 sOpen weights
Read the docs →