​
Overview

Here’s how to configure the models.

​
Migration from Langchain

We have recently switched from Langchain to our own implementation of the models. To migrate the previous code, just replace from langchain_openai import ChatOpenAI with from browser_use.llm import ChatOpenAI etc. The methods should be compatible(ish).

We also made and example here to help you stay with Langchain in case your workflow requires it.

​
Model Recommendations

We recommend using GPT-4.1 for the best performance (best accuracy ~$0.01 per step). The best price to performance can be achieved using gemini-2.0-flash-exp (currently also the most popular model, costs ~$0.001 per step).

​
Supported Models

Our library natively supports the following models:

  • OpenAI
  • Anthropic
  • AWS Bedrock (multiple providers)
  • Azure OpenAI
  • Gemini
  • Groq

We also support all other models that can be called via OpenAI compatible API (deepseek, novita, x, qwen). Please open a PR if you want to add a model.

We have natively switched to structured output when possible,

​
OpenAI

OpenAI’s GPT-4.1 models are recommended for best performance.

from browser_use.llm import ChatOpenAI
from browser_use import Agent

# Initialize the model
llm = ChatOpenAI(
    model="gpt-4.1",
)

# Create agent with the model
agent = Agent(
    task="Your task here",
    llm=llm
)

Required environment variables:

.env
OPENAI_API_KEY=

​
Anthropic

from browser_use.llm import ChatAnthropic
from browser_use import Agent

# Initialize the model
llm = ChatAnthropic(
    model="claude-3-5-sonnet-20240620",
)

# Create agent with the model
agent = Agent(
    task="Your task here",
    llm=llm
)

And add the variable:

.env
ANTHROPIC_API_KEY=

​
Azure OpenAI

from browser_use.llm import ChatAzureOpenAI
from browser_use import Agent
from pydantic import SecretStr
import os

# Initialize the model
llm = ChatAzureOpenAI(
    model="gpt-4.1",
)

# Create agent with the model
agent = Agent(
    task="Your task here",
    llm=llm
)

Required environment variables:

.env
AZURE_OPENAI_ENDPOINT=https://your-endpoint.openai.azure.com/
AZURE_OPENAI_API_KEY=

​
Gemini

[!IMPORTANT] GEMINI_API_KEY was the old environment var name, it should be called GOOGLE_API_KEY as of 2025-05.

from browser_use.llm import ChatGoogle
from browser_use import Agent
from dotenv import load_dotenv

# Read GOOGLE_API_KEY into env
load_dotenv()

# Initialize the model
llm = ChatGoogle(model='gemini-2.0-flash-exp')

# Create agent with the model
agent = Agent(
    task="Your task here",
    llm=llm
)

Required environment variables:

.env
GOOGLE_API_KEY=

​
AWS Bedrock

AWS Bedrock provides access to multiple model providers through a single API. We support both a general AWS Bedrock client and provider-specific convenience classes.

​
General AWS Bedrock (supports all providers)

from browser_use.llm import ChatAWSBedrock
from browser_use import Agent

# Works with any Bedrock model (Anthropic, Meta, AI21, etc.)
llm = ChatAWSBedrock(
    model="anthropic.claude-3-5-sonnet-20240620-v1:0",  # or any Bedrock model
    aws_region="us-east-1",
)

# Create agent with the model
agent = Agent(
    task="Your task here",
    llm=llm
)

​
Anthropic Claude via AWS Bedrock (convenience class)

from browser_use.llm import ChatAnthropicBedrock
from browser_use import Agent

# Anthropic-specific class with Claude defaults
llm = ChatAnthropicBedrock(
    model="anthropic.claude-3-5-sonnet-20240620-v1:0",
    aws_region="us-east-1",
)

# Create agent with the model
agent = Agent(
    task="Your task here",
    llm=llm
)

​
AWS Authentication

Required environment variables:

.env
AWS_ACCESS_KEY_ID=
AWS_SECRET_ACCESS_KEY=
AWS_DEFAULT_REGION=us-east-1

You can also use AWS profiles or IAM roles instead of environment variables. The implementation supports:

  • Environment variables (AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_DEFAULT_REGION)
  • AWS profiles and credential files
  • IAM roles (when running on EC2)
  • Session tokens for temporary credentials
  • AWS SSO authentication (aws_sso_auth=True)

​
Groq

from browser_use.llm import ChatGroq
from browser_use import Agent

llm = ChatGroq(model="meta-llama/llama-4-maverick-17b-128e-instruct")

agent = Agent(
    task="Your task here",
    llm=llm
)

Required environment variables:

.env
GROQ_API_KEY=

​
Ollama

from browser_use.llm import ChatOllama
from browser_use import Agent

llm = ChatOllama(model="llama3.1:8b")