Who contributed to OpenAI on June 8, 2026?

2 developers shipped this update, including jif-oai and rka-oai.

What were the notable OpenAI updates?

feat: add v2 agent residency lru, [codex] Enable standalone web search in code mode, and feat: count V2 concurrency by active execution.

OpenAI

Codex, the SDKs, and the engine behind ChatGPT

Pick a date

Topics: AI / ML Full archive →

The Wire · Showcase

MULTI-AGENT V2 RESIDENCY LAYER COMPLETE

By RepoJournal · Filed 22:42 UTC on June 8, 2026 · About OpenAI

2 people shipped this

jif-oai @jif-oai 4 cited

rka-oai @rka-oai 1 cited

Codex shipped the memory management backbone that separates durable agent identity from thread residency, unblocking true concurrent execution counting across the multi-agent stack.

The v2 agent model just matured past the threading bottleneck. After last week's reload-on-delivery change, agents could survive thread unloads; today's LRU paging [1] moves that separation into `ThreadManager` itself, letting the system page out loaded subagents when sessions get crowded without losing logical identity. That foundation enables the concurrency recount [3] shipping in parallel: v2 now counts active non-root turns instead of resident threads, with admission checks that stay synchronous but allow slight overshot on competing checks. The API surface gets sharper too [4]. Renaming `close_agent` to `interrupt_agent` [4] fixes the contract - v2 agents stay reusable after interruption, so the tool name should match what it does: pause the current turn, not destroy the agent. Web search in code mode [2] also lands today, consuming plaintext output from `/v1/alpha/search` and exposing `web.run` to nested JavaScript calls. Finally, the resume path gets fixed [5]: the old logic still eagerly reopened entire descendant trees, which defeats the point of residency management. Interrupted agents that go idle are now eligible for eviction instead of staying pinned.

Action items

→ Review interrupt_agent semantics in your multi-agent orchestration code - it no longer destroys agent state openai/codex [plan]
→ Test web search integration in code mode before rolling out to production tasks openai/codex [monitor]
→ Verify session residency behavior under load - concurrency admission is now best-effort openai/codex [monitor]

References

[1] feat: add v2 agent residency lru ↗ openai/codex
[2] [codex] Enable standalone web search in code mode ↗ openai/codex
[3] feat: count V2 concurrency by active execution ↗ openai/codex
[4] Rename multi-agent v2 close_agent to interrupt_agent ↗ openai/codex
[5] Avoid reopening v2 descendants on resume ↗ openai/codex

Quick answers

What shipped in OpenAI on June 8, 2026?: Codex shipped the memory management backbone that separates durable agent identity from thread residency, unblocking true concurrent execution counting across the multi-agent stack. In total, 29 commits, 29 pull requests, and 2 releases landed.
Who contributed to OpenAI on June 8, 2026?: 2 developers shipped this update, including jif-oai and rka-oai.
What were the notable OpenAI updates?: feat: add v2 agent residency lru, [codex] Enable standalone web search in code mode, and feat: count V2 concurrency by active execution.

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 9 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 9 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.