OpenAI Python SDK v2.34.0 adds Admin API key support and auth improvements
The latest release of the OpenAI Python client library includes new Admin API key authentication options, expanded user and project parameter support, and multiple authentication bug fixes.
1 source · cross-referenced
- OpenAI released v2.34.0 of its official Python SDK on May 4, 2026, adding support for per-endpoint Admin API keys and bearer-free admin authentication.
- New features include external_key_id for projects and email/metadata parameters for users, along with environment-variable-based header configuration.
- The release resolves seven authentication-related bugs, including Azure auth header handling, type preservation for API keys, and stream helper authentication requirements.
- Performance improvements optimize file structure copying in multipart requests, and documentation updates clarify rate limits for file operations.
OpenAI released v2.34.0 of its Python SDK on May 4, 2026, bringing expanded authentication capabilities and bug fixes. The release is published on the official openai-python GitHub repository, the authoritative source for the library's development.
New authentication features enable per-endpoint Admin API key support and add support for setting HTTP headers via environment variables, improving deployment flexibility for applications that require fine-grained credential scoping. The SDK now handles bearer-free admin authentication, reducing reliance on default fallback mechanisms.
The release addresses seven authentication and type-handling bugs. Fixes include proper preservation of Python API key attribute types, resolution of Azure-specific auth header validation, and corrections to timestamp field types (from float to int) in response models. Stream helper functions now enforce bearer token authentication requirements explicitly.
Additional updates add external_key_id support to project objects and email/metadata parameters to user endpoints, expanding the SDK's coverage of the Admin API surface. Documentation improvements clarify rate limits and vector store information for file creation endpoints. The release also includes a performance optimization for multipart file uploads that reduces memory overhead during file structure copying.
- Oct 13, 2026 · Google DeepMind — Blog
Google DeepMind releases Gemini 4 Argon with domain-leading performance and multi-domain benchmarks
Trust79 - Oct 9, 2026 · Hugging Face
Hugging Face publishes Falcon ASR, a 1.6B-parameter Arabic speech recognition model with Emirati focus
Trust84 - Oct 8, 2026 · OpenAI — News
OpenAI disrupts two AI-enabled influence operations using its models
Trust79