browser-use--browser-use/skills/open-source/references/models.md

Supported LLM Models

Browser Use natively supports 15+ LLM providers. Most providers accept any model string — check each provider's docs to see which models are available.

Quick Reference

Provider	Class	Env Variable
Browser Use Cloud	ChatBrowserUse	BROWSER_USE_API_KEY
OpenAI	ChatOpenAI	OPENAI_API_KEY
Anthropic	ChatAnthropic	ANTHROPIC_API_KEY
Google Gemini	ChatGoogle	GOOGLE_API_KEY
Azure OpenAI	ChatAzureOpenAI	AZURE_OPENAI_*
AWS Bedrock	ChatAWSBedrock	AWS_ACCESS_KEY_ID
DeepSeek	ChatDeepSeek	DEEPSEEK_API_KEY
Mistral	ChatMistral	MISTRAL_API_KEY
Groq	ChatGroq	GROQ_API_KEY
Cerebras	ChatCerebras	CEREBRAS_API_KEY
Ollama	ChatOllama	—
OpenRouter	ChatOpenRouter	OPENROUTER_API_KEY
Vercel AI Gateway	ChatVercel	AI_GATEWAY_API_KEY
OCI (Oracle)	ChatOCIRaw	OCI config file
LiteLLM	ChatLiteLLM	Provider-specific

Recommendations by Use Case

Based on our benchmark of real-world browser tasks:

• Maximum performance: Browser Use Cloud bu-ultra — 78% accuracy, ~14 tasks/hour
• Best open-source + cloud LLM: ChatBrowserUse(model='bu-2-0') — 63.3% accuracy, outperforms every standalone frontier model
• Best standalone model: claude-opus-4-6 — 62% accuracy, excels at custom JavaScript and structured data extraction
• Best value: claude-sonnet-4-6 — 59% accuracy, near-opus quality at lower cost
• Fast + capable: gemini-3-1-pro — 59.3% accuracy

Table of Contents

• Browser Use Cloud (Recommended)
• OpenAI
• Anthropic
• Google Gemini
• Azure OpenAI
• AWS Bedrock
• DeepSeek
• Mistral
• Groq
• Cerebras
• Ollama (Local)
• OpenRouter
• Vercel AI Gateway
• OCI (Oracle)
• LiteLLM (100+ Providers)
• OpenAI-Compatible APIs

---

Browser Use Cloud

Optimized for browser automation — highest accuracy, fastest speed, lowest token cost.

from browser_use import Agent, ChatBrowserUse

llm = ChatBrowserUse()                    # bu-latest (default)
llm = ChatBrowserUse(model='bu-2-0')      # Premium model

Env: BROWSER_USE_API_KEY — get at https://cloud.browser-use.com/new-api-key

Models & Pricing (per 1M tokens):

Model	Input	Cached	Output
bu-1-0 / bu-latest (default)	$0.20	$0.02	$2.00
bu-2-0 (premium)	$0.60	$0.06	$3.50
browser-use/bu-30b-a3b-preview (OSS)	—	—	—

OpenAI

from browser_use import Agent, ChatOpenAI

llm = ChatOpenAI(model="gpt-5")

Env: OPENAI_API_KEY | Available models

Supports custom base_url for OpenAI-compatible APIs.

Anthropic

from browser_use import Agent, ChatAnthropic

llm = ChatAnthropic(model='claude-sonnet-4-6', temperature=0.0)

Env: ANTHROPIC_API_KEY | Available models

Coordinate clicking is automatically enabled for claude-sonnet-4-* and claude-opus-4-* models.

Google Gemini

from browser_use import Agent, ChatGoogle

llm = ChatGoogle(model="gemini-2.5-flash")
llm = ChatGoogle(model="gemini-3-pro-preview")

Env: GOOGLE_API_KEY (free at https://aistudio.google.com/app/u/1/apikey) | Available models

Supports Vertex AI via ChatGoogle(model="...", vertexai=True).

Note: GEMINI_API_KEY is deprecated, use GOOGLE_API_KEY.

Azure OpenAI

Supports the Responses API for codex and computer-use models.

from browser_use import Agent, ChatAzureOpenAI

llm = ChatAzureOpenAI(
    model="gpt-5",
    api_version="2025-03-01-preview",
    azure_endpoint=os.getenv("AZURE_OPENAI_ENDPOINT"),
    api_key=os.getenv("AZURE_OPENAI_API_KEY"),
)

Env: AZURE_OPENAI_ENDPOINT, AZURE_OPENAI_API_KEY | Available models

AWS Bedrock

from browser_use import Agent, ChatAWSBedrock

llm = ChatAWSBedrock(model="us.anthropic.claude-sonnet-4-20250514-v1:0", region="us-east-1")

# Or via Anthropic wrapper
from browser_use import ChatAnthropicBedrock
llm = ChatAnthropicBedrock(model="us.anthropic.claude-sonnet-4-20250514-v1:0", aws_region="us-east-1")

Env: AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_DEFAULT_REGION | Available models

Supports profiles, IAM roles, SSO via standard AWS credential chain. Install with pip install "browser-use[aws]".

DeepSeek

from browser_use import Agent, ChatDeepSeek

llm = ChatDeepSeek(model="deepseek-chat")

Env: DEEPSEEK_API_KEY | Available models

Mistral

from browser_use import Agent, ChatMistral

llm = ChatMistral(model="mistral-large-latest")

Env: MISTRAL_API_KEY | Available models

Groq

from browser_use import Agent, ChatGroq

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

Env: GROQ_API_KEY | Available models

Cerebras

from browser_use import Agent, ChatCerebras

llm = ChatCerebras(model="llama3.3-70b")

Env: CEREBRAS_API_KEY | Available models

Ollama (Local)

from browser_use import Agent, ChatOllama

llm = ChatOllama(model="llama3", num_ctx=32000)

Available models. Requires ollama serve running locally. Use num_ctx for context window (default may be too small).

OpenRouter

Access 300+ models from any provider through a single API.

from browser_use import Agent, ChatOpenRouter

llm = ChatOpenRouter(model="anthropic/claude-sonnet-4-6")

Env: OPENROUTER_API_KEY | Available models

Vercel AI Gateway

Proxy to multiple providers with automatic fallback:

from browser_use import Agent, ChatVercel

llm = ChatVercel(
    model='anthropic/claude-sonnet-4-6',
    provider_options={
        'gateway': {
            'order': ['vertex', 'anthropic'],  # Fallback order
        }
    },
)

Env: AI_GATEWAY_API_KEY (or VERCEL_OIDC_TOKEN on Vercel) | Available models

OCI (Oracle)

from browser_use import Agent, ChatOCIRaw

llm = ChatOCIRaw(
    model="meta.llama-3.1-70b-instruct",
    service_endpoint="https://inference.generativeai.us-chicago-1.oci.oraclecloud.com",
    compartment_id="your-compartment-id",
)

Requires ~/.oci/config setup and pip install "browser-use[oci]". Available models. Auth types: API_KEY, INSTANCE_PRINCIPAL, RESOURCE_PRINCIPAL.

LiteLLM (100+ Providers)

Requires separate install (pip install litellm).

from browser_use.llm.litellm import ChatLiteLLM

llm = ChatLiteLLM(model="openai/gpt-5")
llm = ChatLiteLLM(model="anthropic/claude-sonnet-4-6")

Supports any LiteLLM model string. Useful when you need a provider not covered by the native integrations above.

OpenAI-Compatible APIs

Any provider with an OpenAI-compatible endpoint works via ChatOpenAI:

Qwen (Alibaba)

llm = ChatOpenAI(model="qwen-vl-max", base_url="https://dashscope-intl.aliyuncs.com/compatible-mode/v1")

Env: ALIBABA_CLOUD

ModelScope

llm = ChatOpenAI(model="Qwen/Qwen2.5-VL-72B-Instruct", base_url="https://api-inference.modelscope.cn/v1")

Env: MODELSCOPE_API_KEY

Novita

llm = ChatOpenAI(model="deepseek/deepseek-r1", base_url="https://api.novita.ai/v3/openai")

Env: NOVITA_API_KEY

LangChain

See example at examples/models/langchain.