openai
7.45.1 · 27 Oct 2026Apache-2.036.8M / weekRepositoryHomepagenpmAtom feed
Upgrading openai from 4.38.2 to 7.45.1
292 releases in betweenevery note read
- 12
- changes to make
- 0
- advisories fixed
- 267
- new
- 401
- fixes
Start here
Before you upgrade
Tick each one off as you go- breaking7.44.0
- breaking7.43.0
- breaking7.0.0
- breaking6.0.0
- breaking5.0.0
- breaking4.91.1
- breaking4.49.1
- deprecated7.19.0
- deprecated7.5.0
- deprecated6.17.0
Security
- 7.5.0
azure: prevent deployment path traversal
- 7.5.0
deps: remediate Dependabot and fixture vulnerabilities
- 7.5.0
Prevent prototype pollution in assistant stream deltas
- 7.5.0
security: prevent global Array.prototype pollution in chat completion streams
- 7.5.0
security: prevent multipart upload content-type header injection
- 7.5.0
security: prevent process-wide prototype pollution via assistant stream snapshot IDs
- 7.5.0
security: prevent streaming upload filename/header injection
- 5.0.0
Update next to 14.2.25 for CVE-2025-29927
Release by release
- fixed
api: clarify prompt cache breakpoint TTL behavior
- fixed
vertex: support google-auth-library v11
- fixed
- new
api: add gpt-6.1-terra model identifier
- fixed
deps-dev: bump @cloudflare/workers-types from 5.20261011.1 to 5.20261018.1 in /ecosystem-tests/cloudflare-worker
- fixed
deps-dev: bump @types/node from 26.6.4 to 26.6.5
- fixed
deps-dev: bump fast-uri from 3.1.7 to 3.1.9 in /ecosystem-tests/ts-browser-webpack
- new
- new
api: add delegation controls and prompt cache lifetimes and scheduled automations via
client.beta.agents.automations, including invocation endpoints and theautomationsagent-tool variant - new
agents: reuse action checks between stream events
- new
- fixed
deps-dev: bump follow-redirects from 1.16.0 to 1.16.1 in /ecosystem-tests/vercel-edge
- fixed
deps-dev: bump launch-editor from 2.14.1 to 2.14.2 in /ecosystem-tests/ts-browser-webpack
- fixed
deps: bump next from 15.5.27 to 15.5.28 in /ecosystem-tests/vercel-edge
- fixed
- new
api: add voice deletion
- fixed
agents: avoid repeated ancestor-path hashing
- fixed
api: avoid excessive work when redacting long log values
- new
- new
Add Vertex AI support for the Responses API
- new
- new
api: add 30-minute cache-write usage counts
- new
api: add compaction trigger IDs
- new
api: add custom voice consent management
- fixed
api: align request encoding and response parsing
- new
- new
realtime: add opt-in bounded connection recovery
- new
699 internal lines (docs, tests, tooling) left out.