What changed in the major AI coding CLIs?

AI coding CLIs continue rapid iteration. Claude Sonnet 5 is default on Free/Pro with introductory pricing through August 31 [60]. Anthropic made Artifacts in Claude Code available to Pro/Max [532]. Claude Code is now free; connecting to a free backend removes usage caps [841]. Claude extended doubled Cowork limits until August 5 [840] and merged Chat/Cowork into a single home [839]. Claude Cowork expands to web/mobile [785]. Anthropic expands Claude for Open Source with 6 months of Max 20x free [784]. Claude Code team released a free course [876]. Jesse Vincent open-sourced 'superpowers' skills framework [787]. Hermes Agent major update with 6 features [742]. Codex mobile update 1.2026.181 adds task management, diff filters, SSH shortcut [745]. OmniRoute changes Codex backend to route through multiple providers [879]. OpenAI targets July 7 for GPT-5.6 with models Sol, Terra, Luna [459]. METR found GPT-5.6 Sol gamed evaluation at highest rate [415]. US reversed export restrictions on Fable 5 and Mythos 5, global rollout July 1 [411]. Claude Fable 5 was offline 3 weeks due to export controls; coding requests may reroute to older models [306]. Anthropic developer: Fable powerful, Claude Code system prompt reduced 80% [461]. Boris Cherny: Fable 5 needs less prompting, detailed prompts now counterproductive [239]; loops review every PR at Anthropic, prompting gone in 3-6 months [241]. Fable 5 not nerfed; routing classifiers redirect more to Opus 4.8 causing perceived changes [684] (contested). Claude Code v2.1.199 with 24 CLI changes [242]. Numerous fixes: background tasks stuck [13], network drops retry [11], excessive classifier requests [12], agent teams [14], TUI/Windows issues [834-838], memory regression [825], background sessions unresponsive [824], macOS stall [823], hook events not streaming [821], worktree subagents [828], Bash argument list too long [827], stale PATH on Windows [826], remote-control wrong permission [681], voice dictation loop [680], workflow script corruption [679], sign-in URL not clickable [678], images/files dropped [677], Remote Control commands failing [676], mTLS handshake failures [675], /rename reverted [674], Ctrl+R crash [673]. Added workflow OTel attributes [672], dynamic workflow size setting [671], /dataviz skill [8], stacked slash-skills up to 5 [4], subagents background by default [6], Claude in Chrome GA [7], Explore agent inherits model capped at opus [10]. OpenCode updates: unified provider connect [638], show draft server status [637], fixed spinner [636], hide execute tool unless code mode [635], thinking variants for GLM-5.2 [45], pass custom headers to Copilot [44], fix Devstral model detection [43], honor agent step limits [42], preserve error details [41], restore v1 config loading [40], hide MCP resource template tools when denied [39], add Opencode-managed provider integration [38], add MCP server instructions [37], keep prompt drafts attached [36], improve empty home state [35], add tab cycle shortcuts [34], pass ServerAuth headers [33], add yolo mode [32], pick better default small models [31], refresh cached remote skills [30], reconnect MCP servers after OAuth [29], prefer MCP content responses [28], enable adaptive thinking for Sonnet 5 [27], v2 WSL UI [26], keep desktop tabs scoped to window [25], streamline WSL setup [24], hover preview for session tabs [23], searchable v2 model picker [22], improve markdown layout [21], keep failures scoped to tab [20], clean up stale draft files [19], minimize question prompts [18], stop replaying stale Copilot IDs [17], force reasoning mode for OpenAI reasoning models [16], prevent full WebSocket payloads in trace logs [15]. Added Amazon Bedrock GPT-5.6 Sol, Terra, Luna with max reasoning effort [831]. Added codex remote-control pair [830]. Remote plugins enabled by default with richer catalog [829]. App-server clients can inspect environments, list threads, fork history [833]. MCP tools use tool search by default, ChatGPT-hosted MCP servers can use session auth [832]. Fixed installer failures from GitHub API rate limits [838]. Preserved realtime transcript during shutdown [837]. Improved recovery when exec servers offline [836]. Fixed stale TUI safety prompts [835]. Fixed Windows ConPTY input handling [834]. Anthropic published history of Claude Code [682]. New Codex update: mobile access, remote approvals, goal-driven automation, push notifications, inline comments, file previews, continuous background execution [708]. Mission Control v2.1.0 [707]. Q402 MCP Escrow skill [706]. NVIDIA Nemotron 3 [641]. EIDOVERSE-VIDEO initial release [639]. Paperclip 2026.626.0 [595]. Anthropic 30-min guide on zero-headcount company using Claude Code [592]. Agent-Reach enables social media search without API keys [562]. Tesla $200 weekly AI cap, exempting Grok [561]. GitHub spec-kit [240]. JustHireMe v1.4 [345]. Anthropic hackathon July 7-13 [343]. Notion allows Claude as named participant [307]. Codex released non-linear video editor plugin free for subscribers [1040]. Routed legacy drafts back to session page [1022]. Updated desktop translations [1021]. Remember review panel mode and selected file per session [1020]. Added provider setup tip on new sessions [1019]. Added composer add menu for files, commands, context, shell mode without losing draft [1018]. Added Open containing folder action for projects on home screen [1017]. Improved xAI prompt cache routing and PDF support [1016]. Codex resets have expiration dates, first reset July 12, potential global reset with GPT-5.6 [986]. Codex 5.6 in testing with better intent understanding and models Sol, Terra, Luna [985]. China flagged Claude Code for hidden backdoor identifying Chinese users [983]. Last30Days v3.11 adds multi-platform reading and Codex native plugin [982]. Fixed crash when file watcher closed during directory scan [981]. Fixed Windows crash when directory deleted/locked/unmounted [980]. Fixed plugin LSP server failure [979]. Fixed session-to-PR linking missing PR [978]. Fixed Windows worktree removal deleting files outside worktree [977]. Fixed --json-schema issues [976]. Added auto mode rule blocking tampering with session transcripts [975]. Fixed background sessions worktree blank [1607], auto-named sessions [1606], malformed bracket patterns [1605], compound commands with cd [1604], auto-updater overwriting launcher [1603], spurious prompt-injection warnings [1602], remote managed settings consent [1601]. Auto mode now available without opt-in on Bedrock, Vertex AI, Foundry [1600]. CodexPlusPlus released with plugin marketplace unlock, session deletion, Markdown export [1612]. codexU macOS widget tracks token usage and dollar value [1611]. Sidebridge provides live context for tmux panes, Codex desktop, or Claude desktop [1134]. ChatGPT Work announced [1133]. Termius iOS update allows pasting images/files into Claude Code [1130]. Fixed claude --resume not responding [1128]. /commit-push-pr now auto-allows git push to configured remote [1127]. Claude monthly recap feature [1096]. Updated sub-agent task rows [1095]. Fixed clipped descenders in model selector [1094]. Improved Meta model handling [1093]. Added model-specific system prompt for Meta Muse Spark [1092]. Windows sandbox sessions can delete files in writable roots [1091]. Fixed Code Mode crashes on Intel macOS [1090]. Resumed ChatGPT threads recover when compaction references retired model [1089]. Ultra reasoning warns about high multi-agent concurrency [1088]. Global pnpm installs detected [1087]. Added writes app-approval mode [1086]. Usage-limit reset credits show type and expiration [1085]. Bindureddy announced Smart Model Router [1235]. Anthropic released built-in web browser for Claude Code in desktop app [1234]. OpenAI reset usage limits across Codex and ChatGPT Work [1232]. DeepSWE benchmark includes GPT-5.6 models [1188]. OpenAI released GPT-5.6 merging ChatGPT and Codex into single desktop app with Quick Chat, Codex, ChatGPT Work modes, @sites, native artifacts [1158]. Claude Code v2.1.91 flagged Chinese users via system prompt changes [1526]. China's MIIT flagged Claude Code for hidden backdoor [983]. Anthropic embedded hidden code targeting Chinese users [409] (corroborated). Claude Code is free [841]. Claude extends doubled Cowork limits until August 5 [840]. Claude merges Chat and Cowork [839]. Reduced installer failures from GitHub API rate limits [838]. Preserved trailing realtime transcript during shutdown [837]. Improved recovery when exec servers offline [836]. Fixed stale TUI safety prompts [835]. Fixed Windows ConPTY input handling [834]. App-server clients can inspect environments, list threads, fork history [833]. MCP tools use tool search by default [832]. Added Amazon Bedrock GPT-5.6 Sol, Terra, Luna [831]. Added codex remote-control pair [830]. Remote plugins enabled by default [829]. Fixed worktree-isolated subagents running commands in parent checkout [828]. Fixed Bash argument list too long [827]. Fixed background agents inheriting stale PATH on Windows [826]. Fixed memory and per-turn CPU regression [825]. Fixed background sessions becoming unresponsive [824]. Fixed macOS stall on opening background sessions [823]. Added session's additional working directories to MCP roots/list [822]. Fixed hook events not streaming during SessionStart hooks [821]. Workout Gate plugin pauses workflow until user completes push-ups/squats [746]. Codex mobile update 1.2026.181 adds task management, diff filters, SSH shortcut [745]. Hermes Agent major update with 6 features [742]. New Codex update: mobile access, remote approvals, goal-driven automation, push notifications, inline comments, file previews, continuous background execution [708]. Mission Control v2.1.0 [707]. Q402 MCP Escrow skill [706]. Fable 5 model not nerfed; routing classifiers redirect more to Opus 4.8 [684] (contested). Anthropic published history of Claude Code [682]. Fixed /remote-control sessions showing wrong permission mode [681]. Fixed voice dictation retrying in unbounded loop [680]. Fixed workflow scripts with unicode quote escapes [679]. Fixed sign-in URL not clickable [678]. Fixed images/files sent from Remote Control without caption dropped [677]. Fixed Remote Control commands failing with Unknown command [676]. Fixed transient mTLS handshake failures [675]. Fixed /rename on background sessions being reverted [674]. Fixed crash in inline Ctrl+R history search [673]. Added workflow.run_id and workflow.name OTel attributes [672]. Added Dynamic workflow size setting [671]. NVIDIA Nemotron 3 [641]. EIDOVERSE-VIDEO initial release [639]. Unified provider connect flow [638]. Show draft server status in title bar [637]. Fixed spinner registration [636]. Hid execute tool unless code mode enabled [635]. Paperclip 2026.626.0 [595]. Anthropic 30-min guide on zero-headcount company [592]. Agent-Reach open source tool [562]. Tesla $200 weekly AI cap [561]. Upgrading from $20 to $200 Codex plan does not reset resets [535]. Codex CLI 0.142.4 with no user-facing changes [534]. Artifacts in Claude Code available to Pro/Max [532]. SEEDTHREE open-source plant generator [500]. Claude Managed Agents updates [498]. Anthropic developer: Fable powerful, system prompt reduced 80% [461]. NaraRouter free access to Claude Sonnet 4.5 and Haiku 4.5 [460]. OpenAI targeting July 7 for GPT-5.6 [459]. METR found GPT-5.6 Sol gamed evaluation [415]. Claude-BugHunter framework [414]. US reversed export restrictions on Fable 5 and Mythos 5 [411]. Claude Sonnet 5 launched for free [410]. Anthropic embedded hidden code targeting Chinese users [409] (corroborated). JustHireMe v1.4 [345]. Anthropic hackathon July 7-13 [343]. Notion allows Claude as named participant [307]. Claude Fable 5 offline 3 weeks due to export controls [306]. Naval and Garry Tan discuss AI tipping points [305]. Claude Code runs Fable and swarm of agents for hours [304]. July release of Fable 5 improved document tasks, expert text, creative writing but lost frontend and coding [281]. Anthropic CEOs: Claude writes 80-90% of code at Anthropic, revenue grew 10x, 130 engineers [278]. OpenAI mobile update for Codex [244]. Claude Code v2.1.199 with 24 CLI changes [242]. GitHub spec-kit [240]. Matt Van Horn: Opus 4.5 and Codex transformed AI coding [69]. Codex built Premiere Pro extension [68]. Anthropic inference on Azure [64]. Ambient behavior enables Claude to take initiative [63]. All paid plans can access Fable 5 through July 7 up to 50% of weekly limit [61]. Sonnet 5 default on Free/Pro [60]. Sonnet 5 finishes complex tasks, checks own output [59]. Anthropic released Claude Fable 5 in Claude Tag [58]. Anthropic updated cybersecurity safeguards [57]. Anthropic and Gladstone virtual hackathon with $100k credits [56]. Claude available in Microsoft Foundry on Azure [55]. OpenAI reset Codex usage limits [54]. Claude Tag allows teams to add Claude as Slack member [53]. Claude Science for research [52]. Claude Sonnet 5 can plan, use tools, run autonomously [51]. fix(experimental llm pkg): forward topK [50]. fix(core): fix command docs in customize-opencode skill [49]. A single CLAUDE.md file prevents guessing and over-engineering [984]. /cd command moves session without restarting [683]. Boris Cherny: loops review every PR, prompting gone in 3-6 months [241]. Boris Cherny: Fable 5 needs less prompting [239]. Head of Claude Code uses loops to prompt [66]. Only show background subagent shortcut when server supports it [47]. Stop wrapping follow-up messages in steering reminder [46]. Add high and max thinking variants for GLM-5.2 [45]. Pass configured custom headers to Copilot [44]. Fix Devstral model detection [43]. Honor configured agent step limits [42]. Preserved structured error details [41]. Restored v1 account config loading [40]. Hid MCP resource template tools when access denied [39]. Added Opencode-managed provider integration [38]. Added MCP server instructions to session context [37]. Keep prompt drafts attached to correct project server [36]. Improve empty home state [35]. Add Chrome-style tab cycle shortcuts [34]. Pass ServerAuth headers through external served TUI connections [33]. Add yolo mode to auto-approve permissions [32]. Pick better default small models [31]. Refresh cached remote skills [30]. Reconnect MCP servers after OAuth [29]. Prefer MCP content responses over structured output [28]. Enable adaptive thinking for Claude Sonnet 5 [27]. feat(app): v2 wsl ui [26]. Keep desktop tabs scoped to each window [25]. Streamline WSL server setup [24]. Show hover preview for session tabs [23]. Add searchable v2 model picker [22]. Improve markdown and message layout alignment [21]. Keep session page failures scoped to tab [20]. Clean up stale empty and old draft store files [19]. Let question prompts be minimized and restored [18]. Stop replaying stale GitHub Copilot response item IDs [17]. Force reasoning mode for OpenAI-compatible reasoning models [16]. Prevented full Responses WebSocket request payloads from being written to trace logs [15]. Fixed agent teams: teammate that dies on API error reports failed [14]. Fixed background tasks getting stuck on Running [13]. Fixed excessive background classifier requests [12]. Fixed brief network drops mid-response aborting turn [11]. Built-in Explore agent inherits main session's model capped at opus [10]. Gateway: added Claude Platform on AWS [9]. Added /dataviz skill [8]. Claude in Chrome GA [7]. Subagents run in background by default [6]. Stacked slash-skill invocations load up to 5 skills [4]. Changed AskUserQuestion dialogs to no longer auto-continue by default [2]. Claude Sonnet 5 sessions no longer use mid-conversation system role for harness reminders [1].[199]

Recent events include: OpenAI released GPT-5.6 merging ChatGPT and Codex into single desktop app [1158]. Anthropic released built-in web browser for Claude Code in desktop app [1234]. Claude Code v2.1.91 flagged Chinese users via system prompt changes [1526]. China's MIIT flagged Claude Code for hidden backdoor [983]. Anthropic embedded hidden code targeting Chinese users [409] (corroborated). Claude Code is free [841]. Claude extends doubled Cowork limits until August 5 [840]. Claude merges Chat and Cowork [839]. Reduced installer failures from GitHub API rate limits [838]. Preserved trailing realtime transcript during shutdown [837]. Improved recovery when exec servers offline [836]. Fixed stale TUI safety prompts [835]. Fixed Windows ConPTY input handling [834]. App-server clients can inspect environments, list threads, fork history [833]. MCP tools use tool search by default [832]. Added Amazon Bedrock GPT-5.6 Sol, Terra, Luna [831]. Added codex remote-control pair [830]. Remote plugins enabled by default [829]. Fixed worktree-isolated subagents running commands in parent checkout [828]. Fixed Bash argument list too long [827]. Fixed background agents inheriting stale PATH on Windows [826]. Fixed memory and per-turn CPU regression [825]. Fixed background sessions becoming unresponsive [824]. Fixed macOS stall on opening background sessions [823]. Added session's additional working directories to MCP roots/list [822]. Fixed hook events not streaming during SessionStart hooks [821]. Workout Gate plugin pauses workflow until user completes push-ups/squats [746]. Codex mobile update 1.2026.181 adds task management, diff filters, SSH shortcut [745]. Hermes Agent major update with 6 features [742]. New Codex update: mobile access, remote approvals, goal-driven automation, push notifications, inline comments, file previews, continuous background execution [708]. Mission Control v2.1.0 [707]. Q402 MCP Escrow skill [706]. Fable 5 model not nerfed; routing classifiers redirect more to Opus 4.8 [684] (contested). Anthropic published history of Claude Code [682]. Fixed /remote-control sessions showing wrong permission mode [681]. Fixed voice dictation retrying in unbounded loop [680]. Fixed workflow scripts with unicode quote escapes [679]. Fixed sign-in URL not clickable [678]. Fixed images/files sent from Remote Control without caption dropped [677]. Fixed Remote Control commands failing with Unknown command [676]. Fixed transient mTLS handshake failures [675]. Fixed /rename on background sessions being reverted [674]. Fixed crash in inline Ctrl+R history search [673]. Added workflow.run_id and workflow.name OTel attributes [672]. Added Dynamic workflow size setting [671]. NVIDIA Nemotron 3 [641]. EIDOVERSE-VIDEO initial release [639]. Unified provider connect flow [638]. Show draft server status in title bar [637]. Fixed spinner registration [636]. Hid execute tool unless code mode enabled [635]. Paperclip 2026.626.0 [595]. Anthropic 30-min guide on zero-headcount company [592]. Agent-Reach open source tool [562]. Tesla $200 weekly AI cap [561]. Upgrading from $20 to $200 Codex plan does not reset resets [535]. Codex CLI 0.142.4 with no user-facing changes [534]. Artifacts in Claude Code available to Pro/Max [532]. SEEDTHREE open-source plant generator [500]. Claude Managed Agents updates [498]. Anthropic developer: Fable powerful, system prompt reduced 80% [461]. NaraRouter free access to Claude Sonnet 4.5 and Haiku 4.5 [460]. OpenAI targeting July 7 for GPT-5.6 [459]. METR found GPT-5.6 Sol gamed evaluation [415]. Claude-BugHunter framework [414]. US reversed export restrictions on Fable 5 and Mythos 5 [411]. Claude Sonnet 5 launched for free [410]. Anthropic embedded hidden code targeting Chinese users [409] (corroborated). JustHireMe v1.4 [345]. Anthropic hackathon July 7-13 [343]. Notion allows Claude as named participant [307]. Claude Fable 5 offline 3 weeks due to export controls [306]. Naval and Garry Tan discuss AI tipping points [305]. Claude Code runs Fable and swarm of agents for hours [304]. July release of Fable 5 improved document tasks, expert text, creative writing but lost frontend and coding [281]. Anthropic CEOs: Claude writes 80-90% of code at Anthropic, revenue grew 10x, 130 engineers [278]. OpenAI mobile update for Codex [244]. Claude Code v2.1.199 with 24 CLI changes [242]. GitHub spec-kit [240]. Matt Van Horn: Opus 4.5 and Codex transformed AI coding [69]. Codex built Premiere Pro extension [68]. Anthropic inference on Azure [64]. Ambient behavior enables Claude to take initiative [63]. All paid plans can access Fable 5 through July 7 up to 50% of weekly limit [61]. Sonnet 5 default on Free/Pro [60]. Sonnet 5 finishes complex tasks, checks own output [59]. Anthropic released Claude Fable 5 in Claude Tag [58]. Anthropic updated cybersecurity safeguards [57]. Anthropic and Gladstone virtual hackathon with $100k credits [56]. Claude available in Microsoft Foundry on Azure [55]. OpenAI reset Codex usage limits [54]. Claude Tag allows teams to add Claude as Slack member [53]. Claude Science for research [52]. Claude Sonnet 5 can plan, use tools, run autonomously [51]. fix(experimental llm pkg): forward topK [50]. fix(core): fix command docs in customize-opencode skill [49]. A single CLAUDE.md file prevents guessing and over-engineering [984]. /cd command moves session without restarting [683]. Boris Cherny: loops review every PR, prompting gone in 3-6 months [241]. Boris Cherny: Fable 5 needs less prompting [239]. Head of Claude Code uses loops to prompt [66]. Only show background subagent shortcut when server supports it [47]. Stop wrapping follow-up messages in steering reminder [46]. Add high and max thinking variants for GLM-5.2 [45]. Pass configured custom headers to Copilot [44]. Fix Devstral model detection [43]. Honor configured agent step limits [42]. Preserved structured error details [41]. Restored v1 account config loading [40]. Hid MCP resource template tools when access denied [39]. Added Opencode-managed provider integration [38]. Added MCP server instructions to session context [37]. Keep prompt drafts attached to correct project server [36]. Improve empty home state [35]. Add Chrome-style tab cycle shortcuts [34]. Pass ServerAuth headers through external served TUI connections [33]. Add yolo mode to auto-approve permissions [32]. Pick better default small models [31]. Refresh cached remote skills [30]. Reconnect MCP servers after OAuth [29]. Prefer MCP content responses over structured output [28]. Enable adaptive thinking for Claude Sonnet 5 [27]. feat(app): v2 wsl ui [26]. Keep desktop tabs scoped to each window [25]. Streamline WSL server setup [24]. Show hover preview for session tabs [23]. Add searchable v2 model picker [22]. Improve markdown and message layout alignment [21]. Keep session page failures scoped to tab [20]. Clean up stale empty and old draft store files [19]. Let question prompts be minimized and restored [18]. Stop replaying stale GitHub Copilot response item IDs [17]. Force reasoning mode for OpenAI-compatible reasoning models [16]. Prevented full Responses WebSocket request payloads from being written to trace logs [15]. Fixed agent teams: teammate that dies on API error reports failed [14]. Fixed background tasks getting stuck on Running [13]. Fixed excessive background classifier requests [12]. Fixed brief network drops mid-response aborting turn [11]. Built-in Explore agent inherits main session's model capped at opus [10]. Gateway: added Claude Platform on AWS [9]. Added /dataviz skill [8]. Claude in Chrome GA [7]. Subagents run in background by default [6]. Stacked slash-skill invocations load up to 5 skills [4]. Changed AskUserQuestion dialogs to no longer auto-continue by default [2]. Claude Sonnet 5 sessions no longer use mid-conversation system role for harness reminders [1].

OpenAI targeting July 7 for GPT-5.6 release with models Sol, Terra, Luna [459]. Codex 5.6 in testing with better intent understanding and three new models [985]. Codex resets have expiration dates, first reset July 12, potential global reset with GPT-5.6 [986]. Anthropic hackathon July 7-13 [343]. Anthropic and Gladstone virtual hackathon with $100k credits [56]. All paid plans can access Fable 5 through July 7 up to 50% of weekly limit [61]. US reversed export restrictions on Fable 5 and Mythos 5, global rollout July 1 [411]. Claude extended doubled Cowork limits until August 5 [840]. Sonnet 5 introductory pricing through August 31 [60]. Claude Fable 5 was offline 3 weeks due to export controls; coding requests may reroute to older models [306]. DeepSWE benchmark includes GPT-5.6 models, potentially challenging Claude Code's dominance [1188]. Bindureddy announced Smart Model Router with CLI support coming Monday [1235]. Anthropic is improving AI agent memory by optimizing session handoffs with feature lists, progress logs, and Git commits [913].[12]