SambaNova
Use SambaCloud's OpenAI-compatible inference with Anvia.
SambaCloud exposes an OpenAI-compatible chat completions surface through the SambaNova developer platform, including SambaCloud and SambaStack. In Anvia, configure OpenAIClient with the SambaNova baseUrl, then pass a SambaCloud model id to completionModel(...). SambaNova also publishes a first-party SDK and a custom-checkpoint deployment path on SambaStack.
Create the Client
import { AgentBuilder } from "@anvia/core";
import { OpenAIClient } from "@anvia/openai";
const client = new OpenAIClient({
baseUrl: "https://api.sambanova.ai/v1",
apiKey: process.env.SAMBANOVA_API_KEY,
});
const model = client.completionModel("Meta-Llama-3.3-70B-Instruct");
const agent = new AgentBuilder("support", model)
.instructions("Answer support questions clearly.")
.build();
const response = await agent.prompt("Hello!").send();
console.log(response.output);baseUrl makes Anvia use the OpenAI-compatible chat completion adapter. The model id is the SambaCloud id, not an Anvia-specific alias.
Get the Model List
When SambaCloud exposes a /v1/models endpoint for your account, listModels() returns the model ids available to your key. Because the client was created with baseUrl, the call goes to SambaNova's /models endpoint.
const models = await client.listModels();
console.table(
models.data.map((model) => ({
id: model.id,
name: model.name,
contextLength: model.contextLength,
})),
);Use the id field directly with completionModel(...).
Available Models
SambaCloud hosts a curated open-model catalog optimized for high token throughput on the RDU. Sample ids include:
| Model | Notes |
|---|---|
Meta-Llama-3.3-70B-Instruct | Meta Llama 3.3 70B Instruct |
Meta-Llama-3.1-8B-Instruct | Meta Llama 3.1 8B Instruct |
Meta-Llama-4-Maverick-17B-128E-Instruct-FP8 | Meta Llama 4 Maverick |
Meta-Llama-4-Scout-17B-16E-Instruct | Meta Llama 4 Scout |
DeepSeek-V3.1 | DeepSeek V3.1 |
DeepSeek-R1 | DeepSeek R1 reasoning |
gpt-oss-120b | OpenAI gpt-oss 120B |
Qwen3-32B | Alibaba Qwen 3 32B |
The catalog rotates as new model versions are added. Pick the model id shown in the SambaCloud model catalog at request time.
Notes
- SambaNova API keys are passed as bearer tokens in the
Authorizationheader. - SambaCloud focuses on high-throughput inference on the SN40 and SN50 RDU. Choose it when you need sustained tokens per second on large open models.
- SambaStack is the on-prem option. Custom checkpoints deployed to a SambaStack cluster can be exposed with their own ids — the OpenAI-compatible surface stays the same shape.
- The OpenAI surface is the supported integration path for SambaCloud, but SambaNova also publishes a first-party Python SDK and a custom-checkpoints deployment API for SambaStack. Use those when you need model-bundling features that are not part of the OpenAI shape.
- Model capabilities still depend on the hosted upstream model. Test the specific model id for tool calling, structured output, streaming, and multimodal support before enabling those features.
For current SambaNova API details, see the SambaNova developer guide and the SambaCloud pricing.
