OpenAI's latest SDK update (3.41.0) includes a new `gpt-4o-mini-tts-2025-03-20` model ID,

EXHIBIT 2356 · FLASH

Oct 28, 2026 23:14 UTC pypi openai

OpenAI's latest SDK update (3.41.0) includes a new `gpt-4o-mini-tts-2025-03-20` model ID, a `/v1/realtime/translations` endpoint, and multiple references to `realtime` and `responses` beta connections, indicating potential real-time streaming and response handling features.

READY TO POST:

openai's new sdk (3.41.0) shipped with a `gpt-4o-mini-tts-2025-03-20` model id. could we be getting new text-to-speech models soon?

looks like openai is working on some realtime capabilities with `realtime` and `responses` beta connections in their latest python sdk update. also, a new `/v1/realtime/translations` endpoint. 👀

openai 3.40.0 -> 3.41.0: +130 strings, -743 strings, 51 interesting [model_id] (1) + gpt-4o-mini-tts-2025-03-20 [feature_flag] (4) + py connection = client.beta.realtime.connect(...).enter() # ... connection.close() + py connection = client.beta.responses.connect(...).enter() # ... connection.close() + py connection = client.beta.realtime.connect(...).enter() # ... connection.close() + py connection = client.beta.responses.connect(...).enter() # ... connection.close() [endpoint] (1) + /v1/realtime/translations [sentence] (49) + py connection = client.beta.realtime.connect(...).enter() # ... connection.close() + py connection = client.beta.responses.connect(...).enter() # ... connection.close() + py connection = client.beta.realtime.connect(...).enter() # ... connection.close() + py connection = client.beta.responses.connect(...).enter() # ... connection.close() + .""" tools: List[AssistantTool] """A list of tool enabled on the assistant. There can be a maximum of 128 tools per assistant. Tools can be of types + Add a file to a vector store and poll until processing is complete. `max_wait_seconds` limits polling after upload and attachment; see `poll` for details. + Aggregated billable usage statistics for web search tool calls. + Aggregated usage statistics for image generation tool calls. + Attach a file to the given vector store and wait for it to be processed. `max_wait_seconds` limits polling after attachment; see `poll` for details. + The ID of an existing item in the same Realtime session. + The number of billable web search requests issued. + The type of this item. Always `item_reference`. + Upload a file to the `files` API and then attach it to the given vector store. Note the file will be asynchronously processed (you can use the alte