186 wires and counting

$ follow OpenAI

Keep up with OpenAI in about 3 minutes: what actually shipped — the commits, pull requests, releases, and security advisories that matter.

or

fair warning: these emails are deeply technical. diffs, version numbers, CVEs, benchmark deltas. if that's not your idea of a good read, this isn't your newsletter.

Folds into your digest — weekly by default, monthly if you prefer. Unsubscribe in one click.

$ status

wire 2026-11-13
stories 116

© 2026 RepoJournal Home Showcase How it works Privacy

$ the-wire · showcase

Reasoning input transformations land in both SDKs as Codex 0.168.0 opens a model provider picker

By RepoJournal · Filed · About OpenAI · Composed from the cited sources · methodology

Both official SDKs now surface the optional `input_transformations` array on responses, so the reasoning items the API removed and why are readable in typed code instead of raw JSON, while Codex 0.168.0 ships provider selection, a side-pane TUI and audio attachments.

feat(api): expose reasoning input transformations on responses openai/openai-python

by marcuswood-oai

openai-python v3.50.0 adds the optional `input_transformations` attribute to responses, letting you read the original reasoning item ID and removal reason from each reported transformation. Reading a transformation no longer requires casting the payload loose.

feat(api): type reasoning input transformations on responses openai/openai-node

by marcuswood-oai

openai-node 7.56.0 mirrors the same work with TypeScript types for the optional `input_transformations` array, exposing each transformation's reasoning item ID, action and removal reason. The rest of the release is dev dependency maintenance.

0.168.0 openai/codex

by github-actions[bot]

Codex 0.168.0 exposes model providers under Sessions to Model providers with a default for new chats, and `features.tui_workspace` adds a resizable side conversation you can zoom with `leader z`. Audio-capable models also accept recordings, and Codex skips confirmed rejected ones on later turns while using clip duration to keep more short recordings in context.

Preserve thread input across TUI session resumes openai/codex

by copyberry

Resuming a TUI session now keeps drafts, queued follow-ups and interrupted steer retries attached to their original thread, instead of letting input cross sessions. Submissions whose delivery is still uncertain wait for matching receipts before the queue advances.

feat(api): expand agent, image, Live, and external key support openai/openai-python

by dpiet-oai

The weaker-specified items in the Python PR add API support for agent sessions, SVG output on image edits, Live sessions with safety identifiers and handling of `live.transport.connected` webhooks, and OCI Vault external keys. Separately, Codex now stages Cloud speed selection per conversation alongside pending model and effort, and dedupes overlapping Apps `tools/list` requests across startups...

Quick answers

What shipped in OpenAI on November 13, 2026?
Both official SDKs now surface the optional `input_transformations` array on responses, so the reasoning items the API removed and why are readable in typed code instead of raw JSON, while Codex 0.168.0 ships provider selection, a side-pane TUI and audio attachments. In total, 55 commits, 55 pull requests, and 6 releases landed.
Who contributed to OpenAI on November 13, 2026?
6 developers shipped this update, including github-actions[bot], copyberry, openai-sdks[bot], marcuswood-oai, openai-sdks, and dpiet-oai.
What were the notable OpenAI updates?
feat(api): expose reasoning input transformations on responses, feat(api): type reasoning input transformations on responses, and 0.168.0.