Providers¶

Configure different LLM providers via LangChain.

Supported Providers¶

Provider	Identifier	Current Models (2025)
Google	`google_genai`	gemini-3-flash-preview, gemini-2.5-flash, gemini-2.5-pro
OpenAI	`openai`	gpt-5.2, gpt-5.2-mini, gpt-4.1
Anthropic	`anthropic`	claude-sonnet-4-5, claude-opus-4-6, claude-haiku-4-5
Groq	`groq`	llama-3.3-70b-versatile, llama-3.1-8b-instant, openai/gpt-oss-120b, openai/gpt-oss-20b, groq/compound
Cohere	`cohere`	command-r, command-r-plus
Mistral	`mistral`	mistral-large, mistral-small

Google Gemini (Default)¶

pip install dataframeit[google]
export GOOGLE_API_KEY="your-key"

# Default - no need to specify
result = dataframeit(df, Model, PROMPT, text_column='text')

# Explicit
result = dataframeit(
    df, Model, PROMPT,
    text_column='text',
    provider='google_genai',
    model='gemini-3-flash-preview'
)

# With extra parameters
result = dataframeit(
    df, Model, PROMPT,
    text_column='text',
    provider='google_genai',
    model='gemini-2.5-pro',
    model_kwargs={
        'temperature': 0.2,
        'top_p': 0.9
    }
)

Recommended Models¶

Model	Use	Cost
`gemini-3-flash-preview`	General use, newest	Low
`gemini-2.5-flash`	General use, fast	Low
`gemini-2.5-pro`	Complex tasks, reasoning	Medium

OpenAI¶

pip install dataframeit[openai]
export OPENAI_API_KEY="your-key"

result = dataframeit(
    df, Model, PROMPT,
    text_column='text',
    provider='openai',
    model='gpt-5.2-mini'
)

# With advanced model
result = dataframeit(
    df, Model, PROMPT,
    text_column='text',
    provider='openai',
    model='gpt-5.2',
    model_kwargs={
        'temperature': 0.2
    }
)

Recommended Models¶

Model	Use	Cost
`gpt-5.2-mini`	General use, economical	Low
`gpt-5.2`	Maximum quality	High
`gpt-4.1`	Coding, precise instructions	Medium

Anthropic Claude¶

pip install dataframeit[anthropic]
export ANTHROPIC_API_KEY="your-key"

result = dataframeit(
    df, Model, PROMPT,
    text_column='text',
    provider='anthropic',
    model='claude-sonnet-4-5'
)

# With max_tokens
result = dataframeit(
    df, Model, PROMPT,
    text_column='text',
    provider='anthropic',
    model='claude-opus-4-6',
    model_kwargs={
        'max_tokens': 4096
    }
)

Recommended Models¶

Model	Use	Cost
`claude-sonnet-4-5`	General use, excellent quality	Medium
`claude-opus-4-6`	Maximum quality, agentic	High
`claude-haiku-4-5`	Fast, economical	Low

Groq¶

pip install dataframeit[groq]
export GROQ_API_KEY="your-key"

result = dataframeit(
    df, Model, PROMPT,
    text_column='text',
    provider='groq',
    model='llama-3.3-70b-versatile'
)

# Faster / cheaper model
result = dataframeit(
    df, Model, PROMPT,
    text_column='text',
    provider='groq',
    model='llama-3.1-8b-instant',
    model_kwargs={
        'temperature': 0.2
    }
)

# GPT-OSS for heavier reasoning tasks
result = dataframeit(
    df, Model, PROMPT,
    text_column='text',
    provider='groq',
    model='openai/gpt-oss-120b'
)

Recommended Models¶

Production:

Model	Context	Throughput	Use
`llama-3.3-70b-versatile`	131K	~280 t/s	General use, good quality
`llama-3.1-8b-instant`	131K	~560 t/s	High speed, economical
`openai/gpt-oss-120b`	131K	~500 t/s	Reasoning, complex tasks
`openai/gpt-oss-20b`	131K	~1000 t/s	Faster than 120b, low cost
`groq/compound`	-	~450 t/s	Agentic system with built-in web search and code execution

Preview (may change or be deprecated):

Model	Use
`meta-llama/llama-4-scout-17b-16e-instruct`	Llama 4 Scout, high speed
`qwen/qwen3-32b`	Qwen3, good for reasoning

Availability and free tier

Groq offers a free tier with per-minute request limits per model. The model catalog changes frequently (especially preview models); check console.groq.com/docs/models for the current list and limits.

Cohere¶

pip install langchain-cohere
export COHERE_API_KEY="your-key"

result = dataframeit(
    df, Model, PROMPT,
    text_column='text',
    provider='cohere',
    model='command-r-plus'
)

Mistral¶

pip install langchain-mistralai
export MISTRAL_API_KEY="your-key"

result = dataframeit(
    df, Model, PROMPT,
    text_column='text',
    provider='mistral',
    model='mistral-large-latest'
)

Brazilian region (São Paulo)¶

The providers above use global public endpoints. To serve from Brazil — for latency, data residency or regulatory reasons — use one of the three options below. In all of them, dataframeit forwards model_kwargs straight to LangChain.

Vertex AI (Gemini in `southamerica-east1`)¶

Two variants. The first one needs no extra package install:

# Variant A: uses langchain-google-genai (already a dep of provider 'google_genai')
result = dataframeit(
    df, Model, PROMPT,
    text_column='text',
    provider='google_genai',
    model='gemini-2.5-flash',
    model_kwargs={
        'vertexai': True,
        'project': '<gcp-project-id>',
        'location': 'southamerica-east1',
    },
)

# Variant B: uses langchain-google-vertexai (dedicated provider)
# pip install langchain-google-vertexai
result = dataframeit(
    df, Model, PROMPT,
    text_column='text',
    provider='google_vertexai',
    model='gemini-2.5-flash',
    model_kwargs={
        'project': '<gcp-project-id>',
        'location': 'southamerica-east1',
    },
)

Authentication (any variant):

gcloud auth application-default login
# OR
export GOOGLE_APPLICATION_CREDENTIALS=/path/service-account.json

AWS Bedrock (`sa-east-1`)¶

pip install langchain-aws
aws configure  # or export AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY

result = dataframeit(
    df, Model, PROMPT,
    text_column='text',
    provider='bedrock_converse',
    model='anthropic.claude-3-5-sonnet-20240620-v1:0',
    model_kwargs={'region_name': 'sa-east-1'},
)

Available model IDs in sa-east-1 change — check the Bedrock console first. For cross-region inference, use the us. prefix (e.g. us.anthropic.claude-...) and set region_name to whatever your account has enabled.

For the legacy Bedrock API (non-converse), switch to provider='bedrock' keeping the same model_kwargs. The newer API (bedrock_converse) is recommended for new projects.

Azure OpenAI (Brazil South)¶

pip install langchain-openai
export AZURE_OPENAI_API_KEY="your-key"
export AZURE_OPENAI_ENDPOINT="https://<your-resource>.openai.azure.com/"
export OPENAI_API_VERSION="2025-03-01-preview"

result = dataframeit(
    df, Model, PROMPT,
    text_column='text',
    provider='azure_openai',
    model='gpt-4o',  # or the deployment name
    model_kwargs={'azure_deployment': '<deployment-name>'},
)

The region is encoded in AZURE_OPENAI_ENDPOINT — provision the resource in "Brazil South" via the Azure portal.

The API version (OPENAI_API_VERSION) changes often. Check the latest stable version at aka.ms/azure-openai-api-versions.

Price Comparison (Approximate - 2025)¶

Provider	Model	Input (1M tokens)	Output (1M tokens)
Google	gemini-3-flash-preview	$0.50	$3.00
Google	gemini-2.5-pro	$1.25	$5.00
OpenAI	gpt-5.2-mini	$0.30	$1.20
OpenAI	gpt-5.2	$5.00	$15.00
Anthropic	claude-sonnet-4-5	$3.00	$15.00
Anthropic	claude-haiku-4-5	$1.00	$5.00

Prices change

Check current prices on provider official websites.

Passing API Key Directly¶

If you prefer not to use environment variables:

result = dataframeit(
    df, Model, PROMPT,
    text_column='text',
    provider='openai',
    model='gpt-5.2-mini',
    api_key='sk-...'  # Your key directly
)

Security

Avoid putting API keys directly in code. Prefer environment variables.

Common Parameters (model_kwargs)¶

Parameter	Description	Providers
`temperature`	Creativity (0-1)	All
`top_p`	Nucleus sampling	All
`max_tokens`	Output limit	All

Providers¶

Supported Providers¶

Google Gemini (Default)¶

Recommended Models¶

OpenAI¶

Recommended Models¶

Anthropic Claude¶

Recommended Models¶

Groq¶

Recommended Models¶

Cohere¶

Mistral¶

Brazilian region (São Paulo)¶

Vertex AI (Gemini in southamerica-east1)¶

AWS Bedrock (sa-east-1)¶

Azure OpenAI (Brazil South)¶

Price Comparison (Approximate - 2025)¶

Passing API Key Directly¶

Common Parameters (model_kwargs)¶

Vertex AI (Gemini in `southamerica-east1`)¶

AWS Bedrock (`sa-east-1`)¶