SambaNova Cloud

Model Aggregator

SambaNova Cloud runs open models (DeepSeek, Llama, Qwen) on custom RDU hardware for record-fast token throughput, billed per token.

Models

Prompt Caching

—

Batch Discount

—

Tool Calling

—

Model6

Model	Input Price	Output Price	Cached Input	Context	Modality
DeepSeek-V3.1DeepSeek	$3	$4.5	—	128K	text
DeepSeek-V3.2DeepSeek	$3	$4.5	—	32K	text
gemma-4-31B-itGemma	$0.38	$1.15	—	128K	text
gpt-oss-120bGPT-OSS	$0.22	$0.59	—	128K	text
Meta-Llama-3.3-70B-InstructLlama	$0.6	$1.2	—	128K	text
MiniMax-M2.7MiniMax	$0.6	$2.4	—	192K	text