Cerebras LLM Provider

The Cerebras provider connects Beluga AI to Cerebras’ inference platform, which uses wafer-scale engine (WSE) hardware for extremely fast inference. Cerebras exposes an OpenAI-compatible API, so this provider supports all standard features including streaming, tool calling, and structured output.

Choose Cerebras when you need the fastest possible inference for Llama models. Cerebras’ wafer-scale hardware delivers extremely high tokens-per-second rates, making it particularly suited for latency-critical applications and high-throughput processing of open-source models.

Installation

go get github.com/lookatitude/beluga-ai/llm/providers/cerebras

Configuration

Field	Required	Default	Description
Model	Yes	—	Model ID (e.g. "llama-3.3-70b")
APIKey	Yes	—	Cerebras API key (csk-...)
BaseURL	No	https://api.cerebras.ai/v1	Override API endpoint
Timeout	No	30s	Request timeout

Environment variables:

Variable	Maps to
CEREBRAS_API_KEY	APIKey

Basic Usage

package main

import (
    "context"
    "fmt"
    "log"
    "os"

    "github.com/lookatitude/beluga-ai/config"
    "github.com/lookatitude/beluga-ai/llm"
    "github.com/lookatitude/beluga-ai/schema"
    _ "github.com/lookatitude/beluga-ai/llm/providers/cerebras"
)

func main() {
    model, err := llm.New("cerebras", config.ProviderConfig{
        Model:  "llama-3.3-70b",
        APIKey: os.Getenv("CEREBRAS_API_KEY"),
    })
    if err != nil {
        log.Fatal(err)
    }

    msgs := []schema.Message{
        schema.NewSystemMessage("You are a helpful assistant."),
        schema.NewHumanMessage("What is the capital of France?"),
    }

    resp, err := model.Generate(context.Background(), msgs)
    if err != nil {
        log.Fatal(err)
    }

    fmt.Println(resp.Text())
}

Streaming

for chunk, err := range model.Stream(context.Background(), msgs) {
    if err != nil {
        log.Fatal(err)
    }
    fmt.Print(chunk.Delta)
}
fmt.Println()

Advanced Features

Tool Calling

modelWithTools := model.BindTools(tools)
resp, err := modelWithTools.Generate(ctx, msgs, llm.WithToolChoice(llm.ToolChoiceAuto))

Structured Output

resp, err := model.Generate(ctx, msgs,
    llm.WithResponseFormat(llm.ResponseFormat{Type: "json_object"}),
)

Generation Options

resp, err := model.Generate(ctx, msgs,
    llm.WithTemperature(0.7),
    llm.WithMaxTokens(2048),
    llm.WithTopP(0.9),
    llm.WithStopSequences("END"),
)

Error Handling

resp, err := model.Generate(ctx, msgs)
if err != nil {
    log.Fatal(err)
}

Direct Construction

import "github.com/lookatitude/beluga-ai/llm/providers/cerebras"

model, err := cerebras.New(config.ProviderConfig{
    Model:  "llama-3.3-70b",
    APIKey: os.Getenv("CEREBRAS_API_KEY"),
})

Available Models

Model ID	Description
llama-3.3-70b	Llama 3.3 70B — highest quality available
llama-3.1-8b	Llama 3.1 8B — ultra-fast small model

Refer to Cerebras’ documentation for the latest model list.