OpenAI Assistants API v2 2026: Files, Threads, Code Interpreter

The Assistants API provides managed infrastructure: threads, file search (RAG), code execution, and function calling.

Core Concepts

Assistant: configured AI with instructions, model, tools (save the ID!)

Thread: persistent conversation (one per user, store in your DB)

Run: executing the assistant on a thread

Tools: Code Interpreter, File Search, Functions

Create an Assistant

python
from openai import OpenAI
client = OpenAI()
assistant = client.beta.assistants.create(
    name='Data Analyst',
    instructions='Analyze data with Code Interpreter. Show Python code before results.',
    model='gpt-4o',
    tools=[{'type': 'code_interpreter'}, {'type': 'file_search'}]
)
print(f'Assistant ID: {assistant.id}')  # Save this!

File Search (Managed RAG)

python
import time
with open('handbook.pdf', 'rb') as f:
    file = client.files.create(file=f, purpose='assistants')
vs = client.beta.vector_stores.create(name='Company Knowledge')
client.beta.vector_stores.files.create(vector_store_id=vs.id, file_id=file.id)
while client.beta.vector_stores.retrieve(vs.id).status != 'completed':
    time.sleep(1)
client.beta.assistants.update(
    assistant_id=assistant.id,
    tool_resources={'file_search': {'vector_store_ids': [vs.id]}}
)

Multi-Turn Conversations

python
thread = client.beta.threads.create()
client.beta.threads.messages.create(
    thread_id=thread.id, role='user',
    content='What is our parental leave policy?'
)
run = client.beta.threads.runs.create_and_poll(
    thread_id=thread.id, assistant_id=assistant.id
)
for msg in client.beta.threads.messages.list(thread.id).data:
    if msg.role == 'assistant':
        for c in msg.content:
            if c.type == 'text':
                print(c.text.value)
        break

Code Interpreter: Data Analysis

python
with open('sales.csv', 'rb') as f:
    csv = client.files.create(file=f, purpose='assistants')
thread = client.beta.threads.create()
client.beta.threads.messages.create(
    thread_id=thread.id, role='user',
    content='Analyze data. Find top 5 products and create a bar chart.',
    attachments=[{'file_id': csv.id, 'tools': [{'type': 'code_interpreter'}]}]
)
run = client.beta.threads.runs.create_and_poll(thread_id=thread.id, assistant_id=assistant.id)
for msg in client.beta.threads.messages.list(thread.id).data:
    if msg.role == 'assistant':
        for c in msg.content:
            if c.type == 'image_file':
                data = client.files.content(c.image_file.file_id)
                with open('chart.png', 'wb') as f:
                    f.write(data.read())
                print('Saved: chart.png')

Streaming

python
with client.beta.threads.runs.stream(
    thread_id=thread.id,
    assistant_id=assistant.id,
) as stream:
    for text in stream.text_deltas:
        print(text, end='', flush=True)

Conclusion

The Assistants API is ideal when you want OpenAI managing RAG infrastructure, thread storage, and code execution. Best for rapid prototyping.