What changed in AI tools: week 42, 2026
12–18 Oct 2026 · 42 updates from 25 tools · 2 verified price or limit changes. Pricing and limits first.
Pricing 2
- Kilo offers Cursor customers free Kilo Enterprise for up to a year
Companies paying for Cursor can run Kilo Enterprise free for the rest of that contract, up to one year, if they claim before the end of November.
Source: blog.kilo.aiCurrent Kilo Code plans & limits - Rosalind Academic Plan gives academic labs sponsored access to GPT-Rosalind
Eligible academic life-sciences labs can get sponsored access to GPT-Rosalind and related research tools.
Source: openai.comCurrent ChatGPT plans & limits
Launch 1
- Lovable Growth tools plan and launch ads for your app
Lovable builds a growth plan and can run campaigns on Google, YouTube, Facebook and Instagram from the build chat.
Source: lovable.dev
Model 8
- Krea 3 edits images at native 2K in under 8 seconds for $0.06
The editing model keeps products and characters consistent, takes up to 6 references, and is live on krea.ai and partner APIs.
Source: krea.ai - Copilot drops Claude Opus 4.8, Opus 5 and GPT-5.6 Terra on 16 Nov
Suggested replacements are Opus 5.5 and GPT-6.1 Sol or GPT-6 Luna; individual annual-plan subscribers keep Opus 4.8.
Source: github.blog - Mistral releases Timestral forecasting models
Probabilistic forecasting in four sizes (5M–209M parameters), CPU-efficient and ready for zero-shot use.
Source: mistral.ai - Kimi K3 arrives in Kiro at a 2.4× credit multiplier
Moonshot's open-weight model has a 1M-token context and experimental support for Pro and higher in Kiro's US geography.
Source: kiro.dev - Grok Voice 3.0 launches at the same API price as 2.0
xAI says the speech-to-speech model ranks #1 on Artificial Analysis's Speech-to-Speech Index (86.9) with under-one-second latency.
Source: x.ai - Gemini 4 Argon is generally available in Copilot Pro+, Max, Business and Enterprise
It is billed at provider list pricing under usage-based billing, with a gradual rollout across VS Code, the CLI, the coding agent and other IDEs.
Source: github.blog - Claude Sonnet 4 reaches end of life in Kiro
As of 14 Oct users must pick another model; Sonnet 5.5 is the current Sonnet.
Source: kiro.dev - Gemini 4 Argon rolls out to Google AI Pro and Ultra subscribers
Google's new frontier model reached the Gemini app and developers on 13 Oct, and Antigravity is now built into the Gemini macOS app.
Source: gemini.google
Feature 14
- @GitHub in Slack and Teams now works across multiple repositories
Create issues, open PRs and investigate bugs across repos without switching conversations.
Source: github.blog - Copilot agentic autofix for code scanning alerts is generally available
Copilot explores the codebase, proposes a fix, reruns CodeQL to confirm it closes the alert, then opens a draft PR; requires Code Security or Advanced Security.
Source: github.blog - ElevenLabs Instant Soundtracks turns a video into music and sound effects
The feature generates music and sound effects from a video automatically.
Source: elevenlabs.io - Cursor cloud agents can build shareable canvases; Environments API in beta
Cloud agent environments now support up to 300 repositories, and large ones load 5× faster.
Source: cursor.com - Claude Code Review is generally available for Team and Enterprise
Reviews now finish about 3× faster (6–10 minutes) with per-repo Quick/Standard/Deep levels; eligible Enterprise orgs can claim at least $10,000 in credits by 15 Jan 2027.
Source: claude.com - Lovable Enterprise apps can run in your own AWS account or Azure subscription
Both deployment targets are in private preview.
Source: lovable.dev - Kiro local sandboxing is generally available
OS-enforced limits stop agent-run commands from reaching local files and credentials, in the IDE and CLI.
Source: kiro.dev - Jasper Multiple Audiences targets up to four audiences per generation
One asset can be written for a whole buying committee.
Source: jasper.ai - Copilot's HydraFusion adds a Max option and newer models
The multi-model orchestrator now draws on Opus 5.5, GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna, sets effort per step, and stays in research preview.
Source: github.blog - Kilo MCP server exposes usage, agent profiles and cloud agents to your agent
A remote MCP server puts the Kilo platform API in front of any MCP-compatible client.
Source: blog.kilo.ai - Lovable shows how much usage your free allowances covered
On Free, Pro and Business, usage details now split credits between daily build credits, the daily chat allowance, monthly grants and general credits.
Source: docs.lovable.dev - Lovable Enterprise can keep AI processing in the EU, at higher credit use
EU inference covers building and published apps; some models aren't available and workspaces use more credits while it is on.
Source: lovable.dev - Cursor's Origin opens an app marketplace with 20+ dev-tool integrations
Apps from Cloudflare, Linear, Sentry, Socket and others cover deploys, CI, review, security and monitoring, and the marketplace is open to public submissions.
Source: cursor.com - Amp's Puck is now reachable from ChatGPT and Claude over MCP
Puck MCP exposes two tools, puck and orbs, so you can hand work to Amp and check running orbs from other MCP clients.
Source: ampcode.com
Releases 11
- Claude Code 2.1.299–302: idle compaction; standalone Claude Design closes 14 Dec
Idle 200K+ token chats are now compacted before the one-hour prompt cache expires, and a new notice says standalone Claude Design (claude.ai/design) closes on 14 Dec 2026.
Source: github.com - Replit: Claude Haiku 5.5 for Core and Pro, Bring Your Own Cloud preview
Haiku 5.5 is in Power Mode's model selector, and enterprises can run the full platform or published apps in their own cloud.
Source: docs.replit.com - Codex CLI 0.163: searchable /config, mid-response steering, parallel read-only tools
The release also adds a leader key in /keymap and a /daemon restart, and turns on Daybreak by default for eligible accounts.
Source: github.com - Amp runners: shared machines, worktrees, Mac app runner and remote desktop
Over a month of updates, runners learned to serve many directories, create Git worktrees, be shared with a workspace, start from the macOS app and expose their desktop.
Source: ampcode.com - Zed 1.24: Git tags, Terminal tabs in the Agent Panel
Also better custom Bedrock models and lower memory and CPU use in long agent threads.
Source: zed.dev - Warp adds Kiro CLI and Muse Code to its third-party agent toolbar
October builds bring first-class support for Kiro CLI and Muse Code, Codex cache-write token reporting and a Markdown raw/rendered toggle.
Source: docs.warp.dev - Copilot CLI 1.0.98 adds Gemini 4 Argon support
The release also improves /ask answers (expand, copy, reopen) and fixes command hooks and MCP status sync.
Source: github.com - Tabnine 6.6.8–6.6.11: Jira in the CLI, Gemini cost data fixed
The OpenCode-based CLI can read Jira issues, double-counted Gemini usage and cost data is corrected, and managed models get the latest GPT models.
Source: docs.tabnine.com - HeyGen Voice API: instant clones, a faster Turbo model and public voices
Instant voice clones need no training and are free during preview, and heygen-voice-1-turbo speaks them faster and cheaper; public voices and previews followed.
Source: docs.heygen.com - Notion 3.8: MCP for agents, 150+ connections, a Skills API
External agents connect over MCP with admin controls, Notion agents reach 150+ tools, a Skills API syncs skills to other agents, and version history is upgraded.
Source: notion.so - Devin Desktop 3.10.67: enterprise extension marketplace, faster large folders
Admins can lock extensions to a private marketplace, and the model picker remembers your last variant per model family.
Source: docs.devin.ai
Incident 5
- Grok models in Cursor return capacity errors for about seven hours
Some Grok requests failed with capacity errors from 3:25 pm on 15 Oct to 10:20 pm PDT, when Cursor declared them working normally.
Source: status.cursor.com - OpenAI reports widespread ChatGPT feature degradation
OpenAI's status page flagged widespread feature degradation, with Login among the affected components, before all services recovered.
Source: status.openai.com - Cursor outage hits the IDE, Cloud Agents, Grok Bot and Origin
A widespread issue caused errors across Cursor from 5:22 pm to 8:07 pm PDT on 12 Oct; Cloud Agents were the last to recover.
Source: status.cursor.com - Vercel build and deployment failures last about six and a half hours
Vercel reported more build and deployment failures, then delayed deployment events, until the backlog cleared at 1:11 pm PDT on 12 Oct.
Source: vercel-status.com - GitHub Issues, Pull Requests and Actions fail; notifications delayed for hours
Elevated 500s began at 4:51 am PDT on 12 Oct and core features recovered by 5:41 am, but notification delivery wasn't fully mitigated until 1:16 pm PDT.
Source: githubstatus.com
Other 1
- Character.ai files for Chapter 11 to pursue a court-supervised sale
The platform stays up during the process, but the company says it will be wound down if no workable bid arrives.
Source: blog.character.ai
Verified price & limit changes
-
plan restructureCursor moves into The X Subscription: new Plus $8 / Premium $30 / Super $100 / Ultra $200 plans with a weekly pool
New customers now buy Cursor as part of The X Subscription (Grok + Cursor + X). Usage moves from two monthly pools to one weekly pool shared with Grok; third-party models (Claude, GPT, Gemini) are no longer included and bill at API list price on Premium and above. Legacy Pro ($20), Pro+ ($60), Ultra ($200) continue for existing subscribers.
-
plan restructureSuperGrok folded into The X Subscription (Plus $8, Premium $30, Super $100, Ultra $200)
xAI announced one subscription for Grok, Cursor and X Premium; Premium bundles the benefits of SuperGrok, Cursor Pro and X Premium+ ('previously $90/month combined'). Existing SuperGrok plans continue.