fix: decode SSE incrementally without limiting event size - #3687

Automated Release PR
---


##
[3.4.0](openai/openai-python@v3.3.1...v3.4.0)
(2026-08-25)


### Features

* **api:** Add obfuscation field to ChatCompletionChunk
([openai#3690](openai#3690))
([c7d8e1d](openai@c7d8e1d))
* **api:** add project residency configuration and cost quantity units
([openai#3726](openai#3726))
([bc4f8ef](openai@bc4f8ef))


### Bug Fixes

* **api:** encode Realtime call offers and session configuration
([openai#3736](openai#3736))
([555ac48](openai@555ac48))
* apply consistent origin checks to WebSocket redirects
([openai#3693](openai#3693))
([1b324d0](openai@1b324d0))
* **azure:** encode deployment names consistently
([openai#3683](openai#3683))
([689538d](openai@689538d))
* **azure:** keep provider validation errors value-free
([openai#3691](openai#3691))
([72529c0](openai@72529c0))
* **azure:** resolve one authentication mode
([openai#3689](openai#3689))
([e3d0681](openai@e3d0681))
* compute custom-code summaries from trusted workflow code
([openai#3692](openai#3692))
([2b5868d](openai@2b5868d))
* create upload example fixtures in private directories
([openai#3686](openai#3686))
([f36e6f7](openai@f36e6f7))
* decode SSE incrementally without limiting event size
([openai#3687](openai#3687))
([2598d53](openai@2598d53))
* keep Python SDK diagnostics metadata-only
([openai#3685](openai#3685))
([600aa8d](openai@600aa8d))
* Preserve Azure authentication boundaries across transports
([openai#3684](openai#3684))
([06ef57c](openai@06ef57c))
* preserve the configured TLS hostname
([openai#3694](openai#3694))
([aa5fbc4](openai@aa5fbc4))
* preserve WebSocket send queue byte accounting during flush
([openai#3688](openai#3688))
([96f966d](openai@96f966d))


### Chores

* **api:** Clarify image background docs and preview support
([openai#3703](openai#3703))
([bedb9a7](openai@bedb9a7))
* **api:** document supported image generation models
([openai#3695](openai#3695))
([8edd9ae](openai@8edd9ae))
* **api:** move chat validation tests out of generated code
([openai#3698](openai#3698))
([9d3ba20](openai@9d3ba20))
* **api:** move webhook tests out of generated code
([openai#3700](openai#3700))
([04ecb3c](openai@04ecb3c))
* **api:** remove redundant generated formatting
([openai#3696](openai#3696))
([5ac1e03](openai@5ac1e03))
* **api:** remove redundant generated test edits
([openai#3697](openai#3697))
([351ef84](openai@351ef84))
* **api:** Update SDK generation metadata only
([openai#3716](openai#3716))
([e43b422](openai@e43b422))
* set a ceiling for Python SDK customization
([openai#3714](openai#3714))
([04d5d79](openai@04d5d79))


### Documentation

* **api:** restore generated ChatKit API index
([openai#3705](openai#3705))
([4534106](openai@4534106))
* standardize Python SDK vulnerability disclosure policy
([openai#3642](openai#3642))
([1fc0a21](openai@1fc0a21))


### Refactors

* **api:** isolate audio response format selection
([openai#3701](openai#3701))
([ece4324](openai@ece4324))
* **api:** isolate vector-store polling helpers
([openai#3713](openai#3713))
([a002ef3](openai@a002ef3))
* **api:** move file processing polling into SDK-owned helpers
([openai#3712](openai#3712))
([5f20c51](openai@5f20c51))
* **api:** share embedding response decoding
([openai#3699](openai#3699))
([50de9af](openai@50de9af))
* **api:** share webhook signature verification
([openai#3704](openai#3704))
([e14ac34](openai@e14ac34))

---
This PR was generated with [Release
Please](https://github.com/googleapis/release-please). See
[documentation](https://github.com/googleapis/release-please#release-please).

Co-authored-by: openai-sdks[bot] <284451331+openai-sdks[bot]@users.noreply.github.com>