Launches, pricing changes and shutdowns that shift the meta. Every entry links its source.
A new /auto-mode-setup command reads your project and recent sessions to teach auto mode about your environment. On Pro and Max plans, Claude Code now warns you when resuming an old conversation would eat a noticeable share of your 5-hour usage limit, and offers to start fresh instead. Also fixes rm -r on mounted drives skipping the safety prompt.
As of Oct 19 these models are gone from most Copilot experiences, including Chat, agent mode and completions. GitHub's suggested swaps: GPT-6.1 Sol for GPT-5.5 and GPT-5.4, GPT-6 Luna for GPT-5.4 mini and GPT-5 mini, Gemini 3.8 Flash and Grok 4.7. Enterprise admins may need to enable the replacements in model policies.
A new --disable-client-telemetry flag turns off client telemetry uploads, --trust-folder covers one-off prompt-mode runs, verified skills and plugins install from GitHub install links, and /share opens a full-screen menu for sharing the current session.
GPT-6 models such as gpt-6-sol and gpt-6.1-sol now work on OpenAI and Azure OpenAI-compatible endpoints; before this, every request was rejected. With no model chosen, the Cline provider uses Claude Sonnet 5.5 again.
A new CLAUDE_CODE_AUTO_MODE_MODEL setting chooses which model the client-side auto mode classifier uses, and /config gets a row to turn idle compaction off or back on. The release also warns that standalone Claude Design (claude.ai/design) closes by December 14, 2026.
Core and Pro builders can pick Claude Haiku 5.5 in Power Mode, a day after Replit made OpenAI models the Agent default. Enterprises get Bring Your Own Cloud options (full platform in private preview, published-app hosting and image exports in early access), plus a new Solutions Partners directory.
Cline Cloud sessions run a task in a hosted sandbox on a GitHub repo and branch you pick, keep going after you close your editor, and notify you when they finish. They are rolling out gradually and are turned on in Settings.
A new README banner says Gemini CLI has not been optimized for Gemini 4 Argon and points users to Antigravity for that model. It also restates that unpaid-tier and Google One users were moved to Antigravity CLI on June 18, 2026.
Apps built in Lovable can now call OpenAI's GPT-Live for voice, the latest Claude models, Google's Gemini 4 Argon, Gemini Omni and Lyria, plus Perplexity, Firecrawl, Parallel and Jev, without leaving Lovable. Lovable picks a model for you or you name one in chat.
CEO Amjad Masad said at TechCrunch Disrupt that OpenAI is now Replit's preferred partner, with OpenAI models the default for Replit Agent, citing cost, partnership and performance. Claude models can still be picked, and a person familiar with the Anthropic partnership told TechCrunch it is still in place.
Sending a follow-up while Codex is responding now steers the active turn right away, and interrupting also stops running Code Mode cells and their nested tool calls. The release also adds a searchable /config and parallel read-only tool calls.
GitHub will remove Claude Opus 4.8, Claude Opus 5 and GPT-5.6 Terra from all Copilot experiences on November 16, 2026. Suggested replacements are Claude Opus 5.5 and GPT-6.1 Sol or GPT-6 Luna. Opus 4.8 stays for individual subscribers on annual plans.
In the terminal, a 200K+ token conversation left idle for about 54 minutes is now compacted before its one-hour prompt cache expires, so the next message doesn't reprocess everything. Tune it with CLAUDE_CODE_IDLE_COMPACT_MIN_TOKENS or turn it off with "idleCompaction": false.
Set a goal like 'my first 100 signups' and Lovable builds a growth plan from your app, then helps with SEO fixes, community posts and Google, YouTube, Facebook and Instagram ads. You approve budgets before anything spends, and Lovable adds no extra fee beyond normal chat usage.
HydraFusion, which combines several models into one answer, now draws on Claude Opus 5.5, GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna, tailors reasoning effort per step, and adds a HydraFusion Max option. Both remain in research preview in VS Code, the Copilot app and Copilot CLI.
Bolt's $9/mo Lite plan, built on open-weight 'Forge' models running on Bolt's own hardware, stopped taking sign-ups on Oct 14. Existing Lite members keep their price.
Google's newest frontier model is generally available in Copilot for Pro+, Max, Business and Enterprise users, across VS Code, Visual Studio, Copilot CLI, the coding agent, JetBrains, Xcode and Eclipse. It is billed at provider list price under usage-based billing, with a gradual rollout.
New customers now buy Cursor through The X Subscription (Plus $8, Premium $30, Super $100, Ultra $200), with a weekly usage pool shared with Grok. Third-party models are billed separately at API list price. Existing plans can be kept.
Every yearly Base44 plan from Starter up now includes a free domain for one year (Starter: yearly plans bought from Oct 4, 2026).
OpenAI rolled GPT-6 Astra into Codex starting Sep 3, 2026. Included within the allowance on Pro and the $100 Business plan; Plus gets limited Astra usage with optional credits.
No new users or workspaces since Jun 22, 2026. Existing users can keep working until the shutdown on Mar 22, 2027.
Cognition renamed Windsurf to Devin Desktop via an over-the-air update, adding an Agent Command Center for local, Devin Cloud and ACP-compatible agents. Plans unchanged.
Copilot individual plans moved to usage-based AI credits (Pro 1,500, Pro+ 7,000, Max 20,000 per month). Completions stay unlimited on paid plans.
Free Gemini CLI users are limited to Flash models; Pro models now require a paid Google AI subscription.