gigachat

library0.2.1pypypi✓ verified 86d ago

Python library for GigaChat API by Sberbank. Current version 0.2.1, supports chat completions, embeddings, function calling, structured output (JSON schema), and streaming. Released regularly with minor updates.

llm-agents ai-ml

pip install gigachat

github ↗package ↗homepage ↗

INSTALL

IMPORT

SIG · GIGACHAT

gigachat

llm-agentspythonv0.2.1

Install

3.9s avg

Import

861ms

Disk

31MB

Pass rate

10/ 10

Env Coverage10 / 10

glibc

3.9–3.13

musl

3.9–3.13

Install & Compatibility

Where this runs

tested against v0.2.1 · pip install

no network on importno background threads

Install × environment matrix

Each cell = how many times install + import succeeded across repeated harness runs. Partial = flaky.
glibc = Debian/Ubuntu slim · musl = Alpine Linux

musl

py 3.10–3.95 runs

installs and imports cleanly · install 0.0s · import 0.900s · 32.7MB

glibc

py 3.10–3.95 runs

installs and imports cleanly · install 3.9s · import 0.822s · 32MB

31MB installed

● package 31MB

Code

Verified usage

Verified import paths — ran on the pinned version, not inferred.

GigaChat

✓ from gigachat import GigaChat

Main client class for API interactions.

Messages

✓ from gigachat.models import Messages

✗ from gigachat import Messages

Messages is a model class, not top-level.

ChatCompletion

✓ from gigachat.models import ChatCompletion

✗ from gigachat import ChatCompletion

ChatCompletion is a model, not directly in gigachat namespace.

Initialize GigaChat client with token, send a chat message, and print the response.

from gigachat import GigaChat import os # Use environment variable for token (replace with your actual token) token = os.environ.get('GIGACHAT_TOKEN', '') with GigaChat(credentials=token, verify_ssl_certs=False) as giga: response = giga.chat("Hello, how are you?") print(response.choices[0].message.content)

Debug

Known issues

breakingIn v0.2.0, the synchronous client no longer uses 'requests' library; it uses 'httpx'. Code that relied on requests internals may break.

fix

Update to use httpx-based patterns. The client API remains the same.

affects: >=0.2.0

breakingIn v0.2.0, Pydantic V2 is used instead of V1. Models like Messages, ChatCompletion have changed their internal schema. Serialization with .dict() now returns a dict, not a Model instance.

fix

If you use .dict() or .json() on models, ensure compatibility with Pydantic V2. Use .model_dump() instead of .dict() if needed.

affects: >=0.2.0

deprecatedThe 'user' parameter in GigaChat().chat() is deprecated in favor of passing user_id in the message dict.

fix

Use message format: {'role': 'user', 'content': '...', 'user_id': '...'}

affects: >=0.1.40

gotchaThe client defaults to verify_ssl_certs=False. For production, set verify_ssl_certs=True to avoid security warnings.

fix

Instantiate with GigaChat(credentials=..., verify_ssl_certs=True)

affects: all

Errors

Common errors & fixes

ImportError: cannot import name 'GigaChat' from 'gigachat'

Older versions (pre 0.1.30) used a different import path.

fix

Upgrade gigachat: pip install --upgrade gigachat

gigachat.exceptions.AuthenticationError: ...

Invalid or missing credentials token.

fix

Ensure you provide a valid token as the 'credentials' parameter or set the GIGACHAT_TOKEN environment variable.

AttributeError: 'Messages' object has no attribute 'dict'

Pydantic V2 removed .dict() for models; use .model_dump() instead.

fix

Replace .dict() with .model_dump() for Pydantic models.

Upgrade

Version history

0.2.1latest on PyPI · released Apr 27, 2026

Audit

Dependencies

httpxrequiredHTTP client for API calls

pydanticrequiredData validation and serialization

Agent activity

19 hits · last 30 days

node

OpenAI (training)

Resources

githubhttps://github.com/ai-forever/gigachat ↗

packagehttps://pypi.org/project/gigachat/ ↗

homepagehttps://developers.sber.ru/gigachat ↗