Who contributed to OpenAI on May 30, 2026?

4 developers shipped this update, including aibrahim-oai, sayan-oai, dylan-hurd-oai, and jif-oai.

What were the notable OpenAI updates?

Use inject_if_running for active goal steering (#24924), [codex] Add model tool mode selector, and Show activity for standalone web search calls.

OpenAI

Codex, the SDKs, and the engine behind ChatGPT

Pick a date

Topics: AI / ML Full archive →

The Wire · Showcase

CODEX OVERHAULS GOAL STEERING AND MODEL TOOL SELECTION

By RepoJournal · Filed 06:02 UTC on May 30, 2026 · About OpenAI

4 people shipped this

aibrahim-oai @aibrahim-oai 1 cited

sayan-oai @sayan-oai 1 cited

dylan-hurd-oai @dylan-hurd-oai 1 cited

jif-oai @jif-oai 1 cited

Goal steering now routes through the same injection path as runtime steering, while models gain explicit tool-mode control independent of legacy feature flags.

The core steering infrastructure shifted overnight with a unified injection approach that treats active-turn goal steering as a first-class runtime operation [1]. This lands on top of recent work moving goal steering onto source-labeled context fragments, meaning your model's steering directives now follow the same queuing rules as everything else flowing through the session. Separately, models can now declare their code-execution behavior through catalog metadata rather than relying on client-side feature flags [2], a critical move for heterogeneous deployments where different models need different execution modes. The config layer got cleaner too, with document-handling helpers moved into their own module [5], reducing cognitive load when scanning the config edit state machine. On the observability front, standalone web search calls now emit proper activity telemetry [3], so you'll see search progress in the extension the same way you see it everywhere else. An experimental toggle lets teams opt out of request_user_input at the config level [4], giving you finer control over which tools get input-request behavior.

Action items

→ Review goal steering integration if you instrument Session::inject_if_running openai/codex [plan]
→ Audit model catalog metadata for tool_mode declarations before deploying heterogeneous model fleets openai/codex [plan]
→ Monitor standalone web search activity logging in production openai/codex [monitor]

References

[1] Use inject_if_running for active goal steering (#24924) openai/codex
[2] [codex] Add model tool mode selector ↗ openai/codex
[3] Show activity for standalone web search calls ↗ openai/codex
[4] feat(config) experimental_request_user_input toggle ↗ openai/codex
[5] Move config document helpers into their own module ↗ openai/codex

Quick answers

What shipped in OpenAI on May 30, 2026?: Goal steering now routes through the same injection path as runtime steering, while models gain explicit tool-mode control independent of legacy feature flags. In total, 40 commits and 40 pull requests landed.
Who contributed to OpenAI on May 30, 2026?: 4 developers shipped this update, including aibrahim-oai, sayan-oai, dylan-hurd-oai, and jif-oai.
What were the notable OpenAI updates?: Use inject_if_running for active goal steering (#24924), [codex] Add model tool mode selector, and Show activity for standalone web search calls.

CODEX HARDENS MCP RUNTIME STABILITY ACROSS FIVE CRITICAL FIXES

OpenAI shipped a coordinated hardening of the Model Context Protocol layer across Codex, patching race conditions in thread startup, resource access, and authority management that could leave stale configurations in production.

ai 66 shipped 1-min read

OpenAI 1 day ago

SDK RELEASES LOCK IN PROMPT CACHING AND SPEND CONTROLS

Both Node and Python clients shipped v6.49.0 and v2.48.0 overnight with admin spend limit APIs and relaxed prompt cache validation.

ai 103 shipped 1-min read

OpenAI 2 days ago

CODEX CONSOLIDATES THREAD STARTUP, NODE SDK ADDS REALTIME AND STANDARD SCHEMA SUPPORT

Codex refactored its threading and database layer for consistency while the Node SDK ships four new schema helpers and fixes audio streaming in a single release push.

ai 120 shipped 1-min read

OpenAI 3 days ago

CODEX FIXES WINDOWS EXECUTION BUGS WHILE TIGHTENING SECURITY REVIEW GATES

Codex shipped critical fixes for Windows process management and terminal navigation that were silently breaking execution sessions on Windows machines.

ai 104 shipped 1-min read

Elsewhere on the wire

AI Agents about 10 hours ago

CLAUDE OPUS 5 LANDS ACROSS THE STACK

The newest Anthropic model is now live in langchain, Cline, and llama-index, with native support for extended reasoning and 1M context windows.

ai-agents 28 shipped 1-min read

Local LLMs about 10 hours ago

OLLAMA LANDS LAGUNA SUPPORT AND CRUSHES MEMORY LEAKS WHILE SGLANG HITS V0.5.16 WITH CONFIDENCE-DRIVEN SPECULATIVE DECODING

Ollama shipped three critical performance and reliability fixes for Metal residency and concurrent access patterns, while SGL-Lang released 0.5.16 with a new speculative algorithm hitting 383.7 tok/s on DeepSeek-V4.