--- collection: frontier-ai-labs collection_name: Frontier AI Labs week_start: 2026-06-01 title: Codex expands beyond coding while Claude and OpenAI harden security and cost controls release_count: 58 generated: 2026-07-11 canonical: https://releases.sh/collections/frontier-ai-labs/digest/2026-06-01 --- # Codex expands beyond coding while Claude and OpenAI harden security and cost controls OpenAI rolled out role-specific Codex plugins, interactive Sites, and enterprise admin controls for ChatGPT, while Anthropic stopped billing for Claude refusals without output and added a max_tokens parameter to the advisor tool. Both platforms shipped significant security hardening across Claude Code, ChatGPT Lockdown Mode, and the Claude API. ### Codex becomes a platform for knowledge work OpenAI made its biggest play yet to turn [Codex](/release/rel_PnOJ6zx895oT2AquahBhI-openai-launches-role-specific-codex-plugins-and-interactive-sites) from a developer tool into a general-purpose work platform. Six new role-specific plugins bundle 62 apps and 110 skills for data analytics, creative production, sales, product design, public equity investing, and investment banking — available in preview for Business and Enterprise customers. More notable is the launch of interactive Sites, which lets Codex create and share live websites and apps via URL directly from a conversation. Annotations now work on documents, spreadsheets, and slides, allowing in-place refinement without leaving the chat. The [CLI](/release/rel_aJEFIhzYcmqXtUDK73mSw-codex-cli-0-137-0-adds-enterprise-credit-limits-remote-control-pairing-and-multi) kept pace with enterprise features: monthly credit limits, cloud-managed config bundles, and remote-control pairing for support workflows. A separate [CLI update](/release/rel_fVwRRibEqjC4jZReNOhY2-codex-0-138-0-adds-desktop-handoff-and-improves-image-attachments-with-file) added Desktop handoff on macOS and Windows, plus file-path exposure for local image attachments so follow-up edits are more reliable. Activity insights and shareable [profile cards](/release/rel_N2xAoYFmqjr1qDYySIAr_-codex-adds-activity-insights-and-profile-sharing-cards) arrived in the Codex app, along with improved Computer Use startup readiness. ### Refusal billing and advisor cost controls on Claude Anthropic quietly shipped a pair of API changes that matter for production budgets. The [Claude API](/release/rel_EFLVBx6mBD2_wvBh4FyKX-claude-api-no-longer-charges-for-refusals-without-output) no longer charges for requests that return `stop_reason: "refusal"` without any generated output — a meaningful cost saver for workloads with aggressive guardrails. Separately, the [advisor tool](/release/rel_4e3DfE8TlrEPcrr_m1wbx-claude-platform-adds-max-tokens-parameter-to-advisor-tool) now supports a `max_tokens` parameter to cap the advisor model's output per call, reducing both latency and token spend on tasks that only need short responses. Both changes are live on the API. ### Hardening sessions, permissions, and platforms Security work dominated the week. [Claude Code v2.1.160](/release/rel_sq2Wv3jWMrxMJtei21PaI-claude-code-v2-1-160-hardens-shell-and-config-file-editing-fixes-background) adds prompts before writing to shell startup files (`.zshenv`, `.zlogin`, `.bash_login`) and `~/.config/git/`, and blocks `acceptEdits` from writing build-tool config files that grant code execution. [v2.1.161](/release/rel_LsFfBql1Ri4EkZGvNu3NP-claude-code-v2-1-161-fixes-managed-settings-auth-bypass-and-background-session) fixed a managed-settings auth bypass where `forceLoginOrgUUID` policies were blocking third-party provider sessions alongside the org pin. [v2.1.162](/release/rel_63v4or3_7wLkaSV8gjueF-claude-code-v2-1-162-fixes-permission-rules-mcp-timeouts-and-startup-hangs) fixed WebFetch permission rules not applying to preapproved domains and Windows permission-rule handling with backslashes, while [v2.1.166](/release/rel_Gcnx3E1TkROSYfS6JJe6Y-claude-code-v2-1-166-hardens-cross-session-messaging-and-adds-fallback-models) hardened cross-session messaging so relayed messages no longer carry user authority — plus added a `fallbackModel` setting and retry logic for API failures. Across all these versions, Anthropic fixed dozens of smaller bugs: background session data loss, startup hangs on read-only config directories, and MCP timeout behavior. OpenAI matched the security momentum. [ChatGPT Lockdown Mode](/release/rel_ilOEIOChmRHpmWKeXtpX1-chatgpt-lockdown-mode-now-available-to-all-logged-in-users) is now available to all logged-in users as an opt-in setting, restricting live web browsing, deep research, agent mode, and file downloads to reduce prompt-injection exfiltration risk. For [Enterprise and Edu](/release/rel_CxKgzHiX29U4fQxps5RW5-chatgpt-enterprise-edu-admins-can-now-configure-app-permissions-for-connected) workspaces, admins gained granular App permissions — "Always ask," "Any changes," or "Important actions" — settable per connected app. Plugin sharing in [Codex](/release/rel_c964_LqnZiUccoSUsGkra-chatgpt-enterprise-edu-now-supports-plugin-sharing-in-codex) is now available by default for eligible Enterprise workspaces, with an MDM off-switch for admins. ### Model updates and platform changes Anthropic [deprecated Claude Opus 4.1](/release/rel_8yMkI-UKTPyqAacWYT6nD-claude-opus-4-1-deprecation-announced) with retirement on the API scheduled for August 5, 2026, recommending migration to Opus 4.8. Enterprise Claude plans gained [custom admin roles](/release/rel_UzD1WwDS-aD9gGJ2LdK80-enterprise-plans-can-manage-admin-permissions-with-custom-roles) for billing and privacy access without full ownership. OpenAI introduced [GPT-Rosalind](/release/rel_1xKO6mFGsRPsJUng839LI-openai-introduces-gpt-rosalind-model-update-for-life-sciences-research), a model update combining GPT-5.5's agentic coding with stronger drug-discovery reasoning, alongside the new LifeSciBench benchmark. [ChatGPT memory](/release/rel_9SyecKQN7LP70s8pMzJzj-chatgpt-upgrades-memory-with-automatic-updates-and-increased-capacity-for-plus) now updates automatically to reduce stale context, and Plus/Pro users get twice the memory capacity. The SDKs all gained moderation endpoints: [Python](/release/rel_ReaidwfVQ4hiKTr35zY6w-openai-python-sdk-v2-41-0-adds-moderation-support-to-responses-and-chat) v2.41.0, [Node](/release/rel_tVFThuNxJ8xk29GcSLOQi-openai-node-sdk-v6-42-0-adds-moderation-support-to-responses-and-chat) v6.42.0, and [Go](/release/rel_AP2d1ig_247Tg6jlQXIiA-openai-go-sdk-v3-39-0-adds-moderation-to-responses-and-chat-completions) v3.39.0 all added `responses.moderation` and `chat_completions.moderation` support. Smaller bug-fix releases shipped for the ChatGPT mobile apps and Claude desktop clients. ## Releases covered ### Anthropic - [Claude API no longer charges for refusals without output](https://releases.sh/release/rel_EFLVBx6mBD2_wvBh4FyKX-refusal-with-no-output-no-longer-billed) - [Claude Platform adds max_tokens parameter to advisor tool](https://releases.sh/release/rel_4e3DfE8TlrEPcrr_m1wbx-advisor-tool-output-length-now-configurable-via-max-tokens) - [Claude Code v2.1.160 hardens shell and config file editing; fixes background session data loss](https://releases.sh/release/rel_sq2Wv3jWMrxMJtei21PaI-shell-startup-files-now-prompt-before-write-background-sessions-retain-chat) - [Claude Code v2.1.161 fixes managed-settings auth bypass and background session state corruption](https://releases.sh/release/rel_LsFfBql1Ri4EkZGvNu3NP-managed-settings-auth-policies-now-enforced-background-sessions-use-correct) - [Claude Code v2.1.162 fixes permission rules, MCP timeouts, and startup hangs](https://releases.sh/release/rel_63v4or3_7wLkaSV8gjueF-permission-rules-now-enforced-mcp-sub-1000ms-timeouts-fixed-startup-hangs) - [Claude Code v2.1.166 hardens cross-session messaging and adds fallback models](https://releases.sh/release/rel_Gcnx3E1TkROSYfS6JJe6Y-cross-session-messages-lose-user-authority-fallback-models-added) - [Claude Opus 4.1 deprecation announced](https://releases.sh/release/rel_8yMkI-UKTPyqAacWYT6nD-claude-opus-4-1-deprecation-announced) - [Enterprise plans can manage admin permissions with custom roles](https://releases.sh/release/rel_UzD1WwDS-aD9gGJ2LdK80-enterprise-plans-can-manage-admin-permissions-with-custom-roles) ### OpenAI - [OpenAI launches role-specific Codex plugins and interactive Sites](https://releases.sh/release/rel_PnOJ6zx895oT2AquahBhI-six-role-specific-plugins-sites-preview-for-interactive-web-apps) - [Codex CLI 0.137.0 adds enterprise credit limits, remote-control pairing, and multi-agent runtime persistence](https://releases.sh/release/rel_aJEFIhzYcmqXtUDK73mSw-enterprise-credit-limits-remote-control-pairing-multi-agent-runtime-choice) - [Codex 0.138.0 adds Desktop handoff and improves image attachments with file paths](https://releases.sh/release/rel_fVwRRibEqjC4jZReNOhY2-app-command-hands-off-to-desktop-image-paths-exposed-to-models) - [Codex adds activity insights and profile sharing cards](https://releases.sh/release/rel_N2xAoYFmqjr1qDYySIAr_-activity-insights-and-shareable-profile-cards-added) - [ChatGPT Lockdown Mode now available to all logged-in users](https://releases.sh/release/rel_ilOEIOChmRHpmWKeXtpX1-lockdown-mode-limits-web-access-to-reduce-prompt-injection-risk) - [ChatGPT Enterprise & Edu: Admins can now configure app permissions for connected apps](https://releases.sh/release/rel_CxKgzHiX29U4fQxps5RW5-app-permissions-for-connected-apps-now-configurable-by-admins) - [ChatGPT Enterprise & Edu now supports plugin sharing in Codex](https://releases.sh/release/rel_c964_LqnZiUccoSUsGkra-workspace-plugin-sharing-now-available-in-codex) - [OpenAI introduces GPT-Rosalind model update for life sciences research](https://releases.sh/release/rel_1xKO6mFGsRPsJUng839LI-gpt-rosalind-gains-agentic-coding-lifescibench-benchmark-debuts) - [ChatGPT upgrades memory with automatic updates and increased capacity for Plus and Pro](https://releases.sh/release/rel_9SyecKQN7LP70s8pMzJzj-memory-now-updates-automatically-plus-pro-get-2x-capacity) - [OpenAI Python SDK v2.41.0 adds moderation support to responses and chat completions](https://releases.sh/release/rel_ReaidwfVQ4hiKTr35zY6w-moderation-support-added-to-responses-and-chat-completions) - [OpenAI Node SDK v6.42.0 adds moderation support to responses and chat completions](https://releases.sh/release/rel_tVFThuNxJ8xk29GcSLOQi-moderation-support-added-to-responses-and-chat-completions) - [OpenAI Go SDK v3.39.0 adds moderation to responses and chat completions](https://releases.sh/release/rel_AP2d1ig_247Tg6jlQXIiA-moderation-support-for-responses-and-chat-completions)