LangChain / LangGraph vs LlamaIndex
LangChain / LangGraph and LlamaIndex side by side: what each one is, verified plan pricing and install commands where we have them, and the full review of each.
LangChain / LangGraph
Most popular LLM framework. 100K+ GitHub stars. Chains, RAG, vector stores, tool use. LangGraph adds stateful multi-agent workflows with cycles and persistence.
LlamaIndex
LLM data framework for connecting custom data sources to language models. Best-in-class RAG, data connectors, and query engines. Python and TypeScript.
At a glance
| Field | | |
|---|---|---|
| Category | AI Frameworks | AI Frameworks |
| Website | langchain.com | llamaindex.ai |
| Pricing | LangChain and LangGraph libraries - free and open source (MIT); LangSmith Developer - $0/seat per month: 1 seat, up to 5k base traces/mo, then pay as you go; LangSmith Plus - $39/seat per month: up to 10k base traces/mo, unlimited seats, 1 free small serverless deployment, then pay as you go; LangSmith Enterprise - custom pricing, self-hosted and hybrid deployment, custom SSO, support SLA; Usage beyond plan allowances is metered in LangChain Standard Units (1 LSU = $1) | LlamaIndex framework - free and open source (MIT); LlamaParse Free - 10K credits included; LlamaParse Starter - 40K credits included, pay-as-you-go up to $500/mo; LlamaParse Pro - 400K credits included, pay-as-you-go up to $5,000/mo; LlamaParse Enterprise - contact sales; credits are 1,000 for $1.25 |
| Install | pip install -U langchain or pip install -U langgraph or npm install langchain @langchain/core | pip install llama-index or pip install llama-index-core llama-index-readers-file llama-index-llms-ollama llama-index-embeddings-huggingface |
| Models | Model-agnostic: install a provider package (langchain-openai, langchain-anthropic, Google Gemini, AWS Bedrock, Azure, OpenRouter and many more) and pick a model with init_chat_model, for example init_chat_model("google_genai:gemini-3.7-flash"). New model names work without a LangChain update because provider packages pass them straight to the provider's API. | Out of the box it uses OpenAI gpt-3.5-turbo for generation and text-embedding-ada-002 for embeddings, and the docs recommend passing text-embedding-3-small for new projects. To use another provider, such as Ollama for the LLM or Hugging Face embeddings, install its integration package and pass the model in. |
| Last verified | October 26, 2026 | October 26, 2026 |
In Depth
LangChain / LangGraph
LangChain is the most popular framework for building LLM applications, with over 100K GitHub stars. It provides abstractions for chains (sequential LLM calls), RAG (retrieval-augmented generation with any vector store), tool use, and output parsing. LangGraph extends it with stateful, graph-based workflows - agents that can loop, branch, and persist state across interactions. A more recent addition is 'Deep Agents' for long-running autonomous tasks. LangSmith provides observability and tracing. The ecosystem is massive - integrations with every model provider, vector database, and tool imaginable. I cover LangChain in my AI Agent Frameworks course.
LlamaIndex
LlamaIndex is the go-to framework for connecting your own data to large language models. It provides data connectors (LlamaHub) for ingesting from PDFs, databases, APIs, Notion, Slack, and hundreds of other sources. The indexing layer chunks, embeds, and stores your data in any vector database. Query engines handle retrieval-augmented generation with support for recursive retrieval, sub-question decomposition, and multi-document synthesis. It also includes an agent framework with tool use and multi-step reasoning. Available in both Python and TypeScript (LlamaIndex.TS). If your use case is primarily about making LLMs smarter with your own data rather than building autonomous agents, LlamaIndex is more focused and mature than LangChain for that specific problem.
FAQ
Answers taken from each tool's verified fact sheet, with the date it was last checked.
How much does LangChain / LangGraph cost?
As of October 26, 2026: LangChain and LangGraph libraries - free and open source (MIT); LangSmith Developer - $0/seat per month: 1 seat, up to 5k base traces/mo, then pay as you go; LangSmith Plus - $39/seat per month: up to 10k base traces/mo, unlimited seats, 1 free small serverless deployment, then pay as you go; LangSmith Enterprise - custom pricing, self-hosted and hybrid deployment, custom SSO, support SLA; Usage beyond plan allowances is metered in LangChain Standard Units (1 LSU = $1). Source: LangSmith pricing, Install LangChain (Python), Install LangChain (JavaScript), Install LangGraph, LangChain models, langchain 1.4.0 release, langchain-core 1.7.0 release, langgraph releases, langgraphjs releases, LangChain license (MIT).
How do I install LangChain / LangGraph?
Official install commands as of October 26, 2026: pip install -U langchain or pip install -U langgraph or npm install langchain @langchain/core.
Which models does LangChain / LangGraph use?
As of October 26, 2026: Model-agnostic: install a provider package (langchain-openai, langchain-anthropic, Google Gemini, AWS Bedrock, Azure, OpenRouter and many more) and pick a model with init_chat_model, for example init_chat_model("google_genai:gemini-3.7-flash"). New model names work without a LangChain update because provider packages pass them straight to the provider's API.
How much does LlamaIndex cost?
As of October 26, 2026: LlamaIndex framework - free and open source (MIT); LlamaParse Free - 10K credits included; LlamaParse Starter - 40K credits included, pay-as-you-go up to $500/mo; LlamaParse Pro - 400K credits included, pay-as-you-go up to $5,000/mo; LlamaParse Enterprise - contact sales; credits are 1,000 for $1.25. Source: run-llama/llama_index README, llama_index LICENSE, llama_index releases, LlamaIndex installation and setup, LlamaIndex pricing.
How do I install LlamaIndex?
Official install commands as of October 26, 2026: pip install llama-index or pip install llama-index-core llama-index-readers-file llama-index-llms-ollama llama-index-embeddings-huggingface.
Which models does LlamaIndex use?
As of October 26, 2026: Out of the box it uses OpenAI gpt-3.5-turbo for generation and text-embedding-ada-002 for embeddings, and the docs recommend passing text-embedding-3-small for new projects. To use another provider, such as Ollama for the LLM or Hugging Face embeddings, install its integration package and pass the model in.
