12 Releases published by 1 person

• v1.93.1

  published Jul 7, 2025

• v1.93.2

  published Jul 8, 2025

• v1.93.3

  published Jul 9, 2025

• v1.94.0

  published Jul 10, 2025

• v1.95.0

  published Jul 10, 2025

• v1.95.1

  published Jul 11, 2025

• v1.96.0

  published Jul 15, 2025

• v1.96.1

  published Jul 15, 2025

• v1.97.0

  published Jul 16, 2025

• v1.97.1

  published Jul 22, 2025

• v1.97.2

  published Jul 30, 2025

• v1.98.0

  published Jul 30, 2025

13 Pull requests merged by 2 people

• release: 1.98.0

  #2503 merged Jul 30, 2025

• release: 1.97.2

  #2494 merged Jul 30, 2025

• release: 1.97.1

  #2482 merged Jul 22, 2025

• release: 1.97.0

  #2473 merged Jul 16, 2025

• release: 1.96.1

  #2468 merged Jul 15, 2025

• release: 1.96.0

  #2464 merged Jul 15, 2025

• release: 1.95.1

  #2457 merged Jul 11, 2025

• release: 1.95.0

  #2456 merged Jul 10, 2025

• release: 1.94.0

  #2455 merged Jul 10, 2025

• fix(embeddings): raise error message on missing data

  #2369 merged Jul 10, 2025

• release: 1.93.3

  #2452 merged Jul 9, 2025

• release: 1.93.2

  #2449 merged Jul 8, 2025

• release: 1.93.1

  #2434 merged Jul 7, 2025

16 Pull requests opened by 15 people

• Fix excessive token usage with Unicode text in realtime event serialization

  #2444 opened Jul 4, 2025

• Make ResponseFunctionWebSearch.action optional

  #2450 opened Jul 8, 2025

• Log the remaining retries info at `info` log level

  #2460 opened Jul 12, 2025

• refactor(core): centralize OpenAI usage via call_ai()

  #2461 opened Jul 12, 2025

• Update README.md

  #2465 opened Jul 14, 2025

• Add LLM-friendly documentation

  #2469 opened Jul 15, 2025

• Bearer token provider addition to openaiclient

  #2470 opened Jul 15, 2025

• fix(beta): expose vector_stores in Beta client (fixes #2451)

  #2478 opened Jul 17, 2025

• Add defaults for `type` fields in types.beta.realtime and support open `str` type for session.model

  #2479 opened Jul 18, 2025

• fix: turn detection handled properly

  #2480 opened Jul 18, 2025

• Fix the issue of no logprobs when calling non openai

  #2490 opened Jul 23, 2025

• Improve the log message at INFO level during retry attempts

  #2491 opened Jul 23, 2025

• Remove extra newline in README

  #2492 opened Jul 24, 2025

• Fix optional logprobs field in ResponseTextDeltaEvent model

  #2500 opened Jul 29, 2025

• release: 1.99.0

  #2504 opened Jul 30, 2025

• Doc: Update streaming examples to use gpt-4o instead of gpt-3.5-turbo-instruct

  #2506 opened Aug 4, 2025

41 Issues closed by 2 people

• python sdk prompt_cache_key support?

  #2499 closed Jul 30, 2025

• Batch Endpoint and Request Input Object URL mismatch

  #2497 closed Jul 25, 2025

• Windows Arm64 support

  #2459 closed Jul 22, 2025

• Response API image generation tool number of image paramater: {'error': {'message': "Unknown parameter: 'tools[0].n'.", 'type': 'invalid_request_error', 'param': 'tools[0].n', 'code': 'unknown_parameter'}}

  #2481 closed Jul 22, 2025

• Code interpreter generates files with invalid paths

  #2466 closed Jul 22, 2025

• Using a prompt version in Evals

  #2475 closed Jul 22, 2025

• Make API docs available via MCP or similar

  #2476 closed Jul 22, 2025

• FileSearchToolCall.file_search has empty results

  #1966 closed Jul 10, 2025

• Populate the audio stream with items added to the conversation

  #2262 closed Jul 10, 2025

• Image Example - `base64` and URL pair incomplete code snippet

  #1684 closed Jul 10, 2025

• httpx.RemoteProtocolError: peer closed connection without sending complete message body (incomplete chunked read)

  #2356 closed Jul 10, 2025

• Unable to create json schema assistants in async

  #1869 closed Jul 10, 2025

• Async example in streaming documentation is missing await and async for

  #2297 closed Jul 10, 2025

• vector_stores not exposed in SDK 1.78.0 beta.client despite official release

  #2347 closed Jul 10, 2025

• AzureOpenAI.audio.speech.create(...) rejects instructions parameter

  #2418 closed Jul 10, 2025

• Chat Completion API Rejects Audio Input

  #2454 closed Jul 10, 2025

• Realtime `ResponseOutputItemDoneEvent` model has incorrect type for the `item` field

  #2126 closed Jul 9, 2025

• An item with type 'web_search_call' received as ResponseOutputMessage

  #2448 closed Jul 9, 2025

• Responses API: response.code_interpreter_call_code.delta coming back as type `ResponseAudioDeltaEvent`

  #2382 closed Jul 9, 2025

• Prompt parameter should be supported in responses.parse()

  #2445 closed Jul 7, 2025

• Support for real time audio streaming using chunk transfer encoding for Whisper

  #1025 closed Jul 7, 2025

• trying to use my assistant; one run creates more than one run

  #1727 closed Jul 7, 2025

• How to upload multi-images for description

  #1791 closed Jul 7, 2025

• How to stream responses from Assistants API? The quickstart example doesn't seem to be working

  #1792 closed Jul 7, 2025

• "grit apply openai" stuck at Analyzing Finding files

  #1834 closed Jul 7, 2025

• Commit a7ebc26 disables Open Telemetry's httpx instrumentation in some scenarios

  #1144 closed Jul 7, 2025

• Throw exception if API key ENV vars are empty

  #2092 closed Jul 7, 2025

• Realtime `TurnDetection` model is missing `create_response` field

  #2125 closed Jul 7, 2025

• More intuitive Microphone helper

  #2308 closed Jul 7, 2025

• Completion create params should be a pydantic model

  #2301 closed Jul 7, 2025

• Shared data content when used with FastAPI backgroundtasks and client.beta.threads.runs.create_and_poll

  #1345 closed Jul 7, 2025

• client.audio.transcriptions.create receive different results depending on OS

  #1705 closed Jul 7, 2025

• Threads messages list misinterprets the block type

  #1346 closed Jul 7, 2025

• when "openai migrate" in Mac, get error, need help please

  #1713 closed Jul 7, 2025

• Default values for Usage metrics

  #2278 closed Jul 7, 2025

• feature request: expose verify param of httpx client

  #2359 closed Jul 7, 2025

• "unsupported_file" error when creating vector store with certain plain-text/markdown files

  #2439 closed Jul 7, 2025

• Light Future (LF) – A Light‑Based Multilayer Communication Protocol

  #2443 closed Jul 7, 2025

• Expose all vector store file kwargs in `create` to `create_and_poll` (namely `attributes`)

  #2362 closed Jul 7, 2025

• How to initiate realtime transcription session?

  #2253 closed Jul 7, 2025

• openai migration error

  #1876 closed Jul 7, 2025

12 Issues opened by 12 people

• Please add audio in streaming response for gpt-4o-audio-preview

  #2507 opened Aug 4, 2025

• Realtime API Pydantic Models incomplete for RealtimeResponseStatus type

  #2502 opened Jul 30, 2025

• Make ResponseOutputMessageParam.id genuinely optional and support message payloads without an OpenAI‑generated ID

  #2501 opened Jul 30, 2025

• Edge Case 17

  #2498 opened Jul 27, 2025

• Validation error for ResponseTextDeltaEvent after updating to 1.97.1

  #2489 opened Jul 23, 2025

• Responses API error handling reads error.message, but spec says message is top-level

  #2487 opened Jul 22, 2025

• Incomplete responses are not handled by `parse_response` when using Structured Output

  #2486 opened Jul 22, 2025

• gpt-4o-transcribe client.audio.transcriptions.create fails on all supported audio types (“This model does not support the format you provided.”)

  #2477 opened Jul 17, 2025

• Uploading PDF via Files API and using in Streaming gives 400 bad request

  #2472 opened Jul 16, 2025

• OpenAI API responses.parse with web_search_preview tool returns corrupted JSON with control characters

  #2458 opened Jul 11, 2025

• `client.beta.vector_stores` Not Exposed in Python SDK Despite Documentation (v1.93.2)

  #2451 opened Jul 8, 2025

• Support `rnet` as http client (rust backed)

  #2446 opened Jul 7, 2025

78 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Log number of retries at INFO level

  #2404 commented on Jul 7, 2025 • 0 new comments

• insufficient_quota in _exceptions.py

  #1671 commented on Jul 7, 2025 • 0 new comments

• Functionality to adjust exponential backoff associated with max_retries option

  #1102 commented on Jul 7, 2025 • 0 new comments

• Add Support for New Administration API Endpoints

  #1593 commented on Jul 7, 2025 • 0 new comments

• feature request: proactive client-side rate limiting

  #1579 commented on Jul 7, 2025 • 0 new comments

• Allow logging request body

  #1522 commented on Jul 7, 2025 • 0 new comments

• Add shell auto completion for different shell, like bash, zsh, fish, powershell

  #843 commented on Jul 7, 2025 • 0 new comments

• `httpx.PoolTimeout` occurs frequently with SyncClient

  #821 commented on Jul 7, 2025 • 0 new comments

• ChatCompletions create() doesn't type-check enums as role

  #1300 commented on Jul 7, 2025 • 0 new comments

• Add Structured Outputs support to Assistants stream() and create_and_poll() Functions

  #1698 commented on Jul 7, 2025 • 0 new comments

• Embeddings.create's types should support `input` being a `tuple[str]`

  #1934 commented on Jul 7, 2025 • 0 new comments

• types: Missing type for batch request input and output objects

  #1937 commented on Jul 7, 2025 • 0 new comments

• responses.parse returns an error in streaming mode

  #2305 commented on Jul 10, 2025 • 0 new comments

• How to Configure openai-python for Dynamic Request Forwarding (A -> B -> Provider) with Streaming?

  #2365 commented on Jul 10, 2025 • 0 new comments

• Add Support for API Key Provider

  #1611 commented on Jul 10, 2025 • 0 new comments

• Improve behaviour for read timeouts

  #809 commented on Jul 10, 2025 • 0 new comments

• Migrating to v1: How to create a fine-tuning job and stream using async client?

  #1154 commented on Jul 10, 2025 • 0 new comments

• Apply more fixes for Pydantic schema incompatibilities with OpenAI structured outputs

  #1659 commented on Jul 10, 2025 • 0 new comments

• BaseModel to jsonschema for Structured Outputs

  #2093 commented on Jul 10, 2025 • 0 new comments

• Structured outputs `response_format` requires `strict` function calling JSON Schema?

  #1733 commented on Jul 10, 2025 • 0 new comments

• AsyncHttpxClientWrapper creates task in destructor -> prevents usage in other threads

  #2440 commented on Jul 25, 2025 • 0 new comments

• Excessive token usage with Cyrillic text in Pydantic schema descriptions due to JSON Unicode escaping

  #2428 commented on Jul 25, 2025 • 0 new comments

• Responses API: Wrong chunk name for the MCP Call arguments

  #2412 commented on Jul 25, 2025 • 0 new comments

• UnicodeEncodeError when running OpenAI SDK test code in Jupyter Notebook

  #2409 commented on Jul 25, 2025 • 0 new comments

• Responses API Code Interpreter

  #2396 commented on Jul 25, 2025 • 0 new comments

• BUG : `file_search` annotations are coming up as `ResponseAudioDeltaEvent`

  #2383 commented on Jul 25, 2025 • 0 new comments

• chunking_strategy at vector store level is not being applied to files.

  #2380 commented on Jul 25, 2025 • 0 new comments

• cached_token_details is missing from realtime_response_usage

  #2355 commented on Jul 25, 2025 • 0 new comments

• Output response types are not usable as parameters

  #2323 commented on Jul 25, 2025 • 0 new comments

• openai.Audio.transcribe() fails with BytesIO unless .name is manually set

  #2315 commented on Jul 25, 2025 • 0 new comments

• Azure Open AI Async filter results in None type content part

  #2313 commented on Jul 25, 2025 • 0 new comments

• Invalid Schema Error when Using examples Field in Pydantic Model for OpenAI Python Client

  #2274 commented on Jul 25, 2025 • 0 new comments

• GTP-4.5 has been deleted from the `ChatModel`

  #2256 commented on Jul 25, 2025 • 0 new comments

• Realtime Session Update Configuration

  #2199 commented on Jul 25, 2025 • 0 new comments

• Possible memory leak in `AsyncCompletions.parse()`

  #2146 commented on Jul 25, 2025 • 0 new comments

• EOF error in GPT structured output streaming due to unexpected leading whitespace in JSON schema

  #2082 commented on Jul 25, 2025 • 0 new comments

• Pydantic conversion logic for structured outputs is broken for models containing dictionaries

  #2004 commented on Jul 25, 2025 • 0 new comments

• Assistants API: Unexpected `tool_call` type in `on_tool_call_created`

  #2162 commented on Jul 25, 2025 • 0 new comments

• extra_headers while using as proxy

  #1975 commented on Jul 25, 2025 • 0 new comments

• APIConnectionError raised when using AsyncOpenAI along FastAPI and uvicorn(uvloop)

  #1927 commented on Jul 25, 2025 • 0 new comments

• Azure OpenAI client transcribes or translates speech depending on the deployment used, not the library function

  #1910 commented on Jul 25, 2025 • 0 new comments

• openai migration on Windows WSL error. instructions are wrong

  #1880 commented on Jul 25, 2025 • 0 new comments

• json_schema structured output type not supported in gpt-4o assistants

  #1857 commented on Jul 25, 2025 • 0 new comments

• timeout not being respected by client.beta.threads.runs.create_and_poll (or .OpenAI)

  #1855 commented on Jul 25, 2025 • 0 new comments

• Moderation Endpoint Schema Mismatch for illicit and illicit_violent fields

  #1786 commented on Jul 25, 2025 • 0 new comments

• beta.chat.completions.parse returns unhandled ValidationError

  #1763 commented on Jul 25, 2025 • 0 new comments

• Typing: when stream is completed, delta in ChatCompletionChunk from azure openai is None; should be ChoiceDelta

  #1677 commented on Jul 25, 2025 • 0 new comments

• Cache-control headers are not set when polling agent status

  #1673 commented on Jul 25, 2025 • 0 new comments

• I can no longer upload files to vector store with AzureOpenAI

  #1591 commented on Jul 25, 2025 • 0 new comments

• Assistant with gpt-4o and gpt-4o-mini may call unsupported tool 'browser' and throw exception

  #1574 commented on Jul 25, 2025 • 0 new comments

• 'FileCitation' object has no attribute 'quote'

  #1498 commented on Jul 25, 2025 • 0 new comments

• Inconsistent base_url behavior with module client and OpenAIClient

  #1373 commented on Jul 25, 2025 • 0 new comments

• DEFER_PYDANTIC_BUILD causes issues

  #1306 commented on Jul 25, 2025 • 0 new comments

• At debug log level API requests to OpenAI get logged with all headers, including an API key, in plaintext

  #1196 commented on Jul 25, 2025 • 0 new comments

• `ChatCompletionToolMessageParam` should have an optional `name` property

  #1078 commented on Jul 25, 2025 • 0 new comments

• Pydantic warnings at every base64 embedding call for Async Azure

  #1039 commented on Jul 25, 2025 • 0 new comments

• Function to calculate number of tokens

  #412 commented on Jul 25, 2025 • 0 new comments

• OpenAI CLI Tools for Chat Fine-Tuning

  #622 commented on Jul 25, 2025 • 0 new comments

• Support for File Inputs In Azure OpenAI

  #2300 commented on Jul 31, 2025 • 0 new comments

• Make the trailing / optional at openai.base_url setting

  #780 commented on Jul 10, 2025 • 0 new comments

• tool resource support function

  #1532 commented on Jul 10, 2025 • 0 new comments

• Add shell auto-completion

  #1603 commented on Jul 22, 2025 • 0 new comments

• assistants: fallback type in discriminated unions

  #1615 commented on Jul 10, 2025 • 0 new comments

• Correctly set Accept header to text/event-stream for completion streaming

  #1815 commented on Jul 10, 2025 • 0 new comments

• Tuple support for embeddings

  #1973 commented on Jul 10, 2025 • 0 new comments

• fix: considering default_headers for realtime headers

  #1976 commented on Jul 7, 2025 • 0 new comments

• Add a pure python web app for interacting with the realtime API

  #1989 commented on Jul 10, 2025 • 0 new comments

• BUG FIX: Pydantic conversion logic for structured outputs is broken for models containing dictionaries

  #2003 commented on Jul 10, 2025 • 0 new comments

• Fix the issue where the task was being scheduled in stopped loop

  #2115 commented on Jul 10, 2025 • 0 new comments

• fix(azure): remove unnecessary model parameter and require azure deployment

  #2123 commented on Jul 10, 2025 • 0 new comments

• Fix inconsistent base_url behavior in module client and OpenAIClient

  #2157 commented on Jul 10, 2025 • 0 new comments

• Add https:// prefix to azure_endpoint when missing

  #2166 commented on Jul 10, 2025 • 0 new comments

• fix: handle whitespace-only content in structured output parsing

  #2216 commented on Jul 10, 2025 • 0 new comments

• docs(client): fix incorrect comment about timeout calculation.

  #2252 commented on Aug 2, 2025 • 0 new comments

• stt microphone live example

  #2254 commented on Jul 10, 2025 • 0 new comments

• Fix incorrectly parsing event fieldnames in yielding responses

  #2363 commented on Jul 24, 2025 • 0 new comments

• Add info on garbage collection of async clients

  #2388 commented on Jul 10, 2025 • 0 new comments

• feat(examples): add robust error handling script

  #2415 commented on Jul 10, 2025 • 0 new comments