Inference cloud built on custom RDU dataflow chips, delivering high-throughput hosting of open-weight models.
| Model |
|---|
Llama 4 Maverick 17B 128E InstructOSSLlama-4-Maverick-17B-128E-Instruct |
Llama 3.3 70B InstructOSSMeta-Llama-3.3-70B-Instruct |
Gemma 3 12B ITOSSgemma-3-12b-it |
DeepSeek V3.2OSSDeepSeek-V3.2 |
DeepSeek V3.1OSSDeepSeek-V3.1 |
DeepSeek V3.1 (continuous batching)OSSDeepSeek-V3.1-cb |
MiniMax M2.5OSSMiniMax-M2.5 |
DeepSeek R1 Distill Llama 70BOSSDeepSeek-R1-Distill-Llama-70B |
GPT-OSS 120BOSSgpt-oss-120b |