AI tool changelog

Runway launches Runway Dev plugins for Claude Code, Codex and Cursor

Runway announced Runway Dev plugins for Claude Code, Codex and Cursor on November 2, 2026. The plugin connects your coding agent to your Runway Dev account through the Runway Dev MCP and adds Runway's skills for building with the official SDKs. Per Runway, the agent first reads your account (available projects, the models each project can use with prices and limits, and your credit balance), then builds with the SDKs, Model Routers, Characters for real-time avatars, Recipes and Workflows your team has published as API endpoints, and can look up error details and why a Model Router picked a given model when a task fails.

Install commands from Runway: in Claude Code run /plugin install runway-api@claude-plugins-official; in Codex run codex plugin marketplace add runwayml/skills and then codex plugin add runway-api@runway-dev; in Cursor type /add-plugin runway-api in Agent chat or install from the Cursor Marketplace. You sign in with your Runway Dev account the first time the agent uses it, and no API key is stored in the plugin. Runway Dev is a separate API platform from the Runway app, and the post does not list prices.

Source: Runway: Runway Dev is in Claude Code, Codex and Cursor (Nov 2, 2026)

This week in AI tools: Copilot Fusion, Claude API changes, Replit Cloud Agents and more (Nov 3, 2026) →

ElevenLabs releases SFX v3, a sound effects model that follows multi-part prompts more closely

ElevenLabs released SFX v3 on November 2, 2026. To use it, set model_id to eleven_text_to_sound_v3 on the POST /v1/sound-generation endpoint. ElevenLabs calls it an incremental update that is less likely to clip on high-frequency sounds such as gongs, chimes and kettle whistles.

The model picks the sound's length from the prompt unless you set duration_seconds, and you can raise prompt_influence above its default of 0.3 to follow the prompt more closely. Looping currently works only with eleven_text_to_sound_v2. Because v2 and v3 share one endpoint, ElevenLabs suggests A/B testing your common prompts on both. The blog post does not mention any price change.

Source: ElevenLabs: Introducing SFX v3 for developers (Nov 2, 2026)

This week in AI tools: Copilot Fusion, Claude API changes, Replit Cloud Agents and more (Nov 3, 2026) →

Claude API: message threads beta, 128 MB requests and a Usage and Cost API; Opus 4.8 fast mode retires Nov 30

Message threads (Nov 2, beta): with the message-threads-2026-08-12 beta header, the API keeps a conversation on the server so each later request sends only the new messages, which Anthropic says means smaller requests and a lower time to first token for long conversations and agent loops. A thread expires 5 minutes after its last request and is deleted no more than 1 hour after its last use. It is available on the Claude API, Claude Platform on AWS, Claude Models on Google Cloud (Anthropic managed) and Claude in Microsoft Foundry.

Fast mode retirement (Oct 30): Anthropic deprecated fast mode for Claude Opus 4.8, with removal on November 30, 2026. After that, requests to claude-opus-4-8 with speed set to fast return an error, and Anthropic points users who want fast mode to Claude Opus 5.5.

Usage and Cost Analytics API (Oct 29, beta): new /v1/organizations/analytics/usage and /v1/organizations/analytics/cost endpoints, behind the analytics-usage-cost-2026-08-24 beta header, report token usage and cost for Claude Console and Claude Enterprise organizations, filterable by model, workspace, API key and the user or service account behind each request. The older Usage and Cost API and the related Claude Enterprise Analytics API report endpoints are now legacy but remain supported.

Bigger requests (Oct 28): the maximum request size for the Messages API and Token Counting API on the first-party Claude API rose from 32 MB to 128 MB.

Source: Claude Platform release notes (Oct 28 to Nov 2, 2026)

This week in AI tools: Copilot Fusion, Claude API changes, Replit Cloud Agents and more (Nov 3, 2026) →

Claude Max 20x is now Ultra, and verified Ultra accounts lose the 5-hour limit

On October 30, 2026, Anthropic renamed its two Max plans: Max 5x is now just Max, still $100 a month, and Max 20x is now Ultra, still $200 a month. Existing Max 20x subscribers move to Ultra automatically, with no price change.

Claude's pricing page now lists Max at 5x Pro usage every 5 hours and every week, with $100 a month in Claude Platform API credits, and Ultra at 10x Pro usage every week with no 5-hour limit for verified accounts and $200 a month in API credits. Both include Fable for up to half the weekly limit. Anthropic's help center says unverified Ultra accounts keep a temporary limit until they are verified automatically, and Ultra is monthly billing only.

Sources: Claude pricing · Claude Help Center: What is the Ultra plan? · @claudeai on X (Oct 30, 2026)

AI tool pricing changes this week (Nov 2, 2026) →

Lovable adds project import from other tools, chat-started Deep security scans and persistent memory

On October 29, 2026 Lovable added Import project to Lovable in the + menu on the dashboard. A project built in Replit, Bolt, Base44, v0 or Manus is exported as a ZIP and uploaded, and a website, including one built with Wix, Squarespace or Webflow, is imported from its address. Your first import is free; after that, imports count as standard build usage and consume credits. The same day, you can ask Lovable in the project chat to review your app's security and it starts a Deep scan in the background alongside the Quick scan. The scans are free, but your request and Lovable's explanation of the findings count as standard build usage.

On October 30 Lovable added memory: it remembers the preferences, decisions and facts you state while you build and applies them in later conversations. Project memory belongs to one project and is shared with everyone who can edit it, personal memory applies to every project you build, and both can be reviewed, corrected or disabled in the Memory settings. The Publish dialog now also shows a progress bar and next steps, and on November 1 Lovable added a downloadable software bill of materials (SPDX 2.3 or CycloneDX 1.5) from the Security view, which is free.

Source: Lovable changelog (Oct 29, Oct 30 and Nov 1, 2026 entries)

This week in AI tools: Copilot Fusion, Claude API changes, Replit Cloud Agents and more (Nov 3, 2026) →

ChatGPT Pro and Business now include up to $500 a month in OpenAI API credits

Since October 28, 2026, eligible ChatGPT Pro and Business subscriptions can claim free OpenAI API credits by linking the subscription to an API organization they own. Per OpenAI's docs, Pro ($100 tier) gets $100 a month, Pro ($200) $200 and Pro ($500) $500. Business gets $20 per paid Standard seat or $100 per paid Premium seat, pooled and capped at $500 per workspace per month.

The subscription must be paid and at least seven days old, you must own both the subscription and the API organization with the same email, and unused credits expire at the end of each billing cycle. The first grant is prorated. It mirrors the monthly API credits Anthropic added to Claude Max, Ultra and Team.

Sources: OpenAI API changelog (Oct 28, 2026) · OpenAI docs: API credits for ChatGPT subscribers · ChatGPT pricing

AI tool pricing changes this week (Nov 2, 2026) →

GitHub Enterprise now includes Copilot for every licensed user, billed on usage

GitHub said on October 28, 2026 that enterprises working with a GitHub or Microsoft account team can now get Copilot included with GitHub Enterprise for every licensed user. Instead of buying Copilot seats, usage is metered from the first request against the enterprise budget. Completions and next edit suggestions keep working with no budget; Chat, code review and Copilot in the cloud need available budget.

Users in the enterprise who don't hold a GitHub Enterprise license can get the same access through Copilot Standalone at $10 per user per month. Neither option includes monthly AI credits per seat. Self-serve customers can't switch yet (GitHub says it is opening this in stages), and Copilot Business and Copilot Enterprise remain on sale. Individual plans (Free, Pro $10, Pro+ $39, Max $100) are unchanged.

Sources: GitHub changelog: GitHub Enterprise now includes access to GitHub Copilot (Oct 28, 2026) · GitHub Copilot plans

AI tool pricing changes this week (Nov 2, 2026) →

GitHub Copilot Fusion is generally available: pick Fusion and Copilot routes and orchestrates the models

At GitHub Universe on October 28, 2026, GitHub made Copilot Fusion generally available. It combines Project HydraFusion's multi-model orchestration with Copilot's auto model selection. When you pick Copilot Fusion in the model picker you are choosing a system, not a model: for each prompt it decides which models to use and how, via a single model, a cascade (an efficient model drafts and a quality check decides whether to escalate) or a critique loop (one model drafts, another reviews, the first revises). You choose an effort level: Standard for everyday coding, High for harder tasks, Max for the most demanding work.

GitHub reports that Fusion Max scored 60 in the Artificial Analysis Coding Agent Index, within two points of Claude Code with Opus 5.5 Max, at about half the average API cost per task, about half as many tokens and about 40% less time per task. On billing, you pay for each model Fusion uses at that model's standard rate, and for a limited time users on paid Copilot plans get a 30% discount on model costs when using Fusion. It is available on all Copilot plans in VS Code, the GitHub Copilot app, Copilot CLI and JetBrains IDEs, with more clients coming, and administrators can set Fusion as the default model through Enterprise managed settings.

Sources: GitHub changelog: Route and orchestrate models with GitHub Copilot Fusion (Oct 28, 2026) · GitHub Blog: GitHub Universe 2026, evolving every layer of the stack

This week in AI tools: Copilot Fusion, Claude API changes, Replit Cloud Agents and more (Nov 3, 2026) →

GitHub Copilot app adds multi-repo Projects and desktop computer use; Agent Merge arrives on pull requests

GitHub's Universe releases on October 28, 2026 put several features in the Copilot app. Projects, now in public preview, give work that spans several repositories, or needs specs, designs and assets, one home: you describe a goal, add context such as repositories, design mocks, Slack threads or notes, choose models and tools, and Copilot runs as many agents as needed and brings back results such as pull requests. Computer use, in public preview on macOS and Windows, lets Copilot read app content, click controls and enter text in desktop apps, including ones without an API; Copilot asks for approval first and organizations can disable it through managed settings. You can also generate and edit images in chat.

Separately, Agent Merge, already in the Copilot app and VS Code, is now built into pull requests on GitHub: choose Watch and fix in the merge box and Copilot works in a cloud development environment to fix failing Actions workflows, merge conflicts or both, and it can be combined with auto-merge. It is available to users on a paid Copilot plan, subject to organization and enterprise policies. GitHub also said it is transitioning GitHub Copilot SDK into the Microsoft Copilot SDK, with no immediate changes required for existing users.

Sources: GitHub changelog: GitHub Copilot app, Universe releases (Oct 28, 2026) · GitHub changelog: Agent Merge watches and fixes pull requests in the cloud (Oct 28, 2026) · GitHub changelog: GitHub Copilot SDK is becoming Microsoft Copilot SDK (Oct 28, 2026)

This week in AI tools: Copilot Fusion, Claude API changes, Replit Cloud Agents and more (Nov 3, 2026) →

Suno opens an MCP beta so Claude, ChatGPT and Claude Code can make music for you

Suno announced Suno MCP (beta) on October 28, 2026. Once Suno is connected to Claude, ChatGPT, Claude Code or another MCP client, you can describe a song in plain words and work out lyrics and style with your AI, upload a voice memo or loop to turn into a full song, and remix songs from your library into covers, extensions and mashups. Every song lands in your Suno library and can be added to a playlist from the chat.

Suno says the beta is available on all Suno plans and that songs use your credits. Access is by early-access sign-up per Suno's release note.

Source: Suno release notes: Suno MCP (beta) (Oct 28, 2026)

This week in AI tools: Copilot Fusion, Claude API changes, Replit Cloud Agents and more (Nov 3, 2026) →

Replit Cloud Agents open pull requests against your company's existing repositories

Replit announced on October 27, 2026 (updated October 29) that Cloud Agents for existing repositories are available to all Pro and Enterprise customers. Anyone on a team can describe a change in plain language; the agent finds the code, makes the change with a live preview beside the chat, and opens a pull request on GitHub or a merge request on GitLab that follows the team's skills and contribution guidelines. Before the request reaches anyone, the agent tests the change in an isolated cloud sandbox by clicking through the running app and attaches screenshots and video, and it applies review feedback by updating the request.

Replit also describes shared environments that an admin sets up once from the main branch, starting changes by mentioning or messaging @Replit in Slack (available to all customers; Microsoft Teams support is coming), and a self-improving workspace memory that proposes updates to a team's skills for engineers to approve, which is in early access. Replit is holding a webinar on November 5. The post does not state new plan prices.

Source: Replit: Anyone can now bring their ideas into the company codebase (Oct 27, 2026, updated Oct 29)

This week in AI tools: Copilot Fusion, Claude API changes, Replit Cloud Agents and more (Nov 3, 2026) →

OpenAI releases GPT-6.1 Terra at $1 / $5 per 1M tokens, and it is already in GitHub Copilot

OpenAI released GPT-6.1 Terra on October 26, 2026. It is available to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex, and in the API as gpt-6.1-terra (Responses and Chat Completions; tool calling needs the Responses API). Standard API pricing is $1 per 1M input tokens, $0.05 cached input and $5 output, with a 1.05M-token context window and 128K max output. Prompts over 272K input tokens cost 2x input and 1.5x output, and Batch and Flex are 50% off.

OpenAI says Terra's API prices are 50% lower than GPT-5.6 Terra and GPT-6.1 Sol (listed at $2 / $10), and that the same ChatGPT Work and Codex allowance covers more tasks with Terra than with GPT-5.6 Terra, GPT-6.1 Sol or GPT-6 Astra. It still recommends GPT-6.1 Sol and GPT-6 Astra for the most complex coding tasks.

GitHub made GPT-6.1 Terra generally available in Copilot the same day for Pro, Pro+, Max, Business and Enterprise, billed at provider list pricing under usage-based billing, with a gradual rollout across VS Code, Visual Studio, Copilot CLI, the cloud agent, JetBrains, Xcode and Eclipse. OpenRouter also lists it as openai/gpt-6.1-terra.

Sources: OpenAI: Introducing GPT-6.1 Terra (Oct 26, 2026) · OpenAI API docs: GPT-6.1 Terra model page · GitHub changelog: GPT-6.1 Terra in GitHub Copilot · GitHub Docs: Models and pricing for GitHub Copilot

This week in AI tools: GPT-6.1 Terra, Kimi K3.1 Pro and more →

Runway launches App Studio: describe a creative app and Runway builds and hosts it

Runway announced App Studio on October 26, 2026 (its changelog lists the feature from October 23). You describe the tool you need, and App Studio writes, tests, deploys and hosts an app with its own interface that calls Runway's image, video and audio models along with your Brand Kits, Workflows, Assets and References. An existing Agent session or Workflow can be turned into an app in one click, and built-in apps such as Character Swap, AutoComp, Product Portfolio and Precise Edit can be used as-is or remixed.

App Studio is included in every paid plan with no add-on. Building, running and hosting apps draw from your Runway credit balance, and visitors to a publicly shared app pay for their own generations with their own Runway account. Free users can use built-in and shared apps but need a paid plan to build their own. Enterprise admins can turn it off per workspace and control public links.

Sources: Runway: Introducing Runway App Studio (Oct 26, 2026) · Runway changelog

Chroma moves from A to B and Aider from B to C on ToolTier

Chroma joined Replit, and Chroma Cloud stops accepting writes on November 22 and reads on December 22, 2026. The open-source database continues, so Chroma stays a good prototyping pick, but it is no longer an easy managed choice. We moved it from A to B.

Aider has had no release since August 2025 and no pushes since May 2026. It is still a solid git-aware terminal pair programmer, but newer agentic coding CLIs have moved past it, so we moved it from B to C.

Sources: Chroma Cloud wind-down timeline · Aider releases on GitHub

AutoGen moves from C to D on ToolTier after Microsoft archived it

Microsoft archived the AutoGen repository on October 16, 2026. Its README says the project gets no bug fixes, security fixes, new features or support, and Microsoft advises against production use.

After reviewing it, we moved AutoGen from C to D tier. It stays listed with its archived status note for reference. For new projects, look at Microsoft Agent Framework (Microsoft's stated successor), Claude Agent SDK, LangGraph or OpenAI Agents SDK.

Sources: microsoft/autogen README (archived Oct 16, 2026) · AutoGen to Microsoft Agent Framework migration guide

Best AI Agent Frameworks in 2026 →

Moonshot releases Kimi K3.1 Pro, with open weights due by Nov 4; K2 models start retiring

Moonshot AI launched Kimi K3.1 Pro and a faster K3.1 Pro Highspeed variant on October 23, 2026, on Kimi.ai, the Kimi apps, Kimi Work, Kimi Code and the Kimi API. Moonshot says full model weights will be released by November 4, 2026. It reports a 65.7% score on Terminal-Bench 4.0 (its own evaluation with a Claude Code harness).

The Kimi API pricing page lists kimi-k3.1-pro at $2 per 1M input tokens, $10 output and $0.20 cached input, and kimi-k3.1-pro-highspeed at $4 / $20. The model always thinks (reasoning effort low, high or max) and is unlocked after a minimum $1 top-up.

The same pricing page says K2 series models will no longer be available to newly registered users from October 31, 2026, and that Moonshot plans to retire all K2 models by the end of December 2026. Existing users are not affected until then.

Sources: Kimi blog: Introducing Kimi K3.1 Pro · Kimi API docs: Kimi K3.1 Pro quickstart · Kimi API docs: Model inference pricing · Kimi Forum: Kimi K3.1 Pro has launched (Oct 23, 2026)

This week in AI tools: GPT-6.1 Terra, Kimi K3.1 Pro and more →

Chroma is joining Replit and shutting down Chroma Cloud by Dec 22

Chroma announced on October 23, 2026 that it has signed a definitive agreement to join Replit, where the team will keep working on ChromaDB and its Foundation context product. As part of the move, Chroma is winding down its hosted Chroma Cloud service over 60 days.

Per Chroma's wind-down timeline, writes to Chroma Cloud are disabled on November 22, 2026, and reads stop on December 22, 2026. Customers are asked to move write workloads first, then finish exporting data and remove Cloud credentials before the final date. The open-source ChromaDB project stays under the Apache 2.0 license, and Replit says it will keep contributing to it.

If you run Chroma Cloud in production, plan a migration to self-hosted ChromaDB or another vector database. Our Chroma page lists same-category alternatives. Chroma's tier is unchanged while we review it.

Sources: Chroma blog: Chroma is joining Replit (Oct 23, 2026) · Chroma Cloud wind-down timeline

AI tool pricing changes this week (Oct 26, 2026) →

Greptile drops per-seat pricing: Pro is now $1 prepaid credits shared across the org

Greptile announced on October 22, 2026 that it is going fully usage-based, removing per-seat fees in favor of prepaid review credits shared across the team. Its pricing page now lists Pro at $1 per credit with unlimited developers and agents, and the starter plan stays free for one developer with 50 credits a month.

Per Greptile's billing docs, a Base review costs 1 credit, a Plus review 3 and an Apex review 10, and only completed reviews spend credits. When credits run out, new reviews pause until you top up or turn on automatic recharge. Pre-Series A startups can apply for 500 free credits a month for six months, and qualified non-commercial open-source projects can use Greptile free.

Sources: Greptile pricing · Greptile docs: Billing · Greptile announcement on LinkedIn (Oct 22, 2026)

AI tool pricing changes this week (Oct 26, 2026) →

Anthropic donates the Agent Skills standard to the Linux Foundation's Agentic AI Foundation

Anthropic announced on October 22, 2026 that it is donating Agent Skills to the Agentic AI Foundation (AAIF) under the Linux Foundation, the same home it gave the Model Context Protocol. The donation covers the Agent Skills name and logo, the GitHub organization and the agentskills.io domain. Engineers from Anthropic, Google, OpenAI and other companies will maintain the standard as individuals, and you can keep using or contributing to it without joining the AAIF.

A skill is a folder with a SKILL.md file plus any scripts or reference material an agent needs for a task. Anthropic says more than 70 agent platforms, including ChatGPT, Cursor, GitHub Copilot and VS Code, have integrated the standard since December 2025. The proposed roadmap starts with versioned releases, a formal proposal process and security guidance for people who write and install skills. The AAIF, which already hosts MCP and AGENTS.md, announced the same day that it is also taking on OpenAI's Open Responses spec, Hugging Face's OpenEnv and Alibaba's OpenSandbox.

Sources: Anthropic: Donating Agent Skills to the Agentic AI Foundation (Oct 22, 2026) · AAIF: Agentic AI Foundation announces four new projects

AGENTS.md vs CLAUDE.md vs Cursor rules: which file each agent reads →

Anthropic opens a research preview of Claude Duet, a proactive agent app for Pro and Max

Anthropic opened a waitlisted research preview of Claude Duet on October 21, 2026, starting with Pro and Max subscribers in the US. Duet is built around Claude Opus 5.5 and runs as its own app on iOS and Android and on claude.ai. It sends a morning brief across your connected apps, checks them every few hours, and works on tasks in a virtual machine with its own browser, so your laptop or phone doesn't need to stay on.

Duet asks for approval before it sends a message, makes a payment, signs a document or deletes anything, and Anthropic says some actions, such as trading securities or moving money between accounts, are off-limits even with approval. During the preview it is included in Pro and Max and shares your plan's usage limits; its background checks count toward those limits.

Sources: Claude: Introducing Claude Duet · Claude product announcements

Microsoft archives AutoGen and points everyone to Microsoft Agent Framework

Microsoft updated the AutoGen repository on October 16, 2026 to say the project has been archived and is no longer actively maintained. It will get no bug fixes, security fixes, new features or support, and Microsoft advises against production use. The repository stays up for reference.

New users are told to start with Microsoft Agent Framework, which Microsoft describes as AutoGen's production-ready successor (1.0, stable APIs, long-term support), and existing users get a migration guide. On ToolTier, AutoGen's status note is updated; after review we moved it from C to D on Oct 24.

Sources: microsoft/autogen README (updated Oct 16, 2026) · AutoGen to Microsoft Agent Framework migration guide

Best AI Agent Frameworks in 2026 →

ChatGPT gets Slides, and GPT-6 Pro can run several agents in parallel

ChatGPT Slides creates presentations from a conversation, files or an outline, lets teammates edit the same deck in real time, and exports to PowerPoint, PDF or Google Slides on the web. OpenAI says it is rolling out to eligible Pro, Business and Enterprise users with Space access, on the web and in the desktop app.

The same day, GPT-6 Pro in the chat tab got faster responses, better use of plugins and tools, and multiple agents that can work on different parts of a task in parallel, rolling out to eligible Pro users.

Source: OpenAI release notes (Oct 16, 2026)

OpenAI ships GPT-Live-1.1 in ChatGPT and prompt-to-voice in the API

On Oct 15 OpenAI began rolling out GPT-Live-1.1 to all users of ChatGPT, Codex and dots, with 13 new voices, better performance in French, German, Hindi, Mexican Spanish, Korean, Japanese and other languages, and more natural delegation. OpenAI said API access was coming soon.

On Oct 16 the API gained prompt-to-voice: describe a voice in text, then reuse it in GPT-Live sessions. Creating a prompt-generated voice is free; speech uses normal GPT-Live session billing, which OpenAI's pricing page lists at $0.05 per minute for gpt-live-1 (billed per second, backend model and tool usage extra). Nine more natural voices were added for German, French, Hindi, Japanese and Korean.

Sources: OpenAI API changelog · OpenAI community announcement: GPT-Live-1.1 · OpenAI API pricing

ElevenLabs launches Instant Soundtracks: music and synced sound effects for any video

Instant Soundtracks analyzes uploaded footage for sound-effect cues, generates effects aligned to those timestamps and drops them on the timeline as separate layers next to an original score. It can also use the video's existing audio (genre, tempo, instrumentation) as a reference for the new music, and shifts mood across scenes.

It is available in ElevenCreative's Music Video Editor, Creative Engine, the ElevenMusic platform and the ElevenCreative Studio API (new add_sfx, separate_tracks and use_audio_as_reference parameters), with up to 10 videos per request, 200 MB and 10 minutes of video combined.

Source: ElevenLabs blog: Introducing Instant Soundtracks

Cursor opens the Origin App Marketplace, with CodeRabbit and Greptile at launch

Launch apps include Cloudflare, Netlify, Render, Spacelift and Vercel for deployments; Buildkite, CircleCI and others for CI; CodeRabbit, Greptile and Tenki for code review; Semgrep, Socket Security and others for security; Sentry for monitoring; and Linear for project management.

Any team building developer tools can now publish an Origin app: build against the Origin API (app JWTs, installation tokens, webhooks, check runs and PR reviews), make it public and submit it for review. The same day, the Origin API added pull request auto-merge endpoints in preview.

Sources: Cursor changelog: Origin App Marketplace · Origin API changelog (Oct 13, 2026)

Google releases Gemini 4 Argon, its new frontier model, at an introductory API price

Google says Argon is built for long, multi-step work in coding, finance, legal, cybersecurity and science. In the Gemini API (model ID gemini-4-argon) it is paid-tier only: $2 per 1M input tokens and $10 per 1M output tokens (thinking tokens included) through Nov 30, 2026, then $4 and $20 from Dec 1. Context caching is implicit only, at $0.10 per 1M rising to $0.20.

Consumers get it in the Gemini app on Google AI Pro and Ultra, and Ultra subscribers can use it in Google Antigravity. GitHub made it generally available in Copilot the same day for Pro+, Max, Business and Enterprise, billed at provider list pricing. Gemini CLI's README says the CLI has not been optimized for Argon and points users to Antigravity instead.

Sources: Google blog: Gemini 4 Argon is now available for everyone · Gemini API pricing · GitHub changelog: Gemini 4 Argon in GitHub Copilot · Gemini CLI README

Gemini 4 Argon pricing: every way to use it for coding →

Harness acquires Augment Code's Cosmos, Auggie CLI and Context Engine

Augment Code announced on October 8, 2026 that Cosmos, the Auggie CLI, the Code Context Engine and related technology are joining Harness, along with the team that built them. Harness says Cosmos becomes the Harness Cosmos Software Factory Agent inside its Autonomous SDLC Platform, and that customers can keep using their preferred coding tools alongside it.

augmentcode.com is now run by Harness and leads with Cosmos. Its pricing page still lists Standard at $20/month and Business at $100/month (flat team prices with that much usage included), so we have left the plan data as is and added a status note. Augment's tier is unchanged while we watch how the product evolves under Harness.

Sources: Augment Code blog: Augment Code's next chapter (Oct 8, 2026) · Harness blog: Harness acquires Augment Code · Augment Code pricing

OpenAI is retiring gpt-4o-mini-tts and tts-1 in January, and whisper-1 in February 2027

On October 1, 2026, OpenAI deprecated its text-to-speech models tts-1, tts-1-hd and both gpt-4o-mini-tts snapshots, with removal from the API on January 6, 2027. The recommended replacement is gpt-realtime-2.1-mini, which OpenAI's pricing page lists at $10 / $20 per 1M audio input / output tokens.

Earlier, on August 26, 2026, OpenAI told developers that whisper-1, gpt-4o-transcribe, gpt-4o-mini-transcribe and gpt-4o-transcribe-diarize will be removed on February 26, 2027, along with the /v1/audio/translations endpoint. It recommends gpt-live-transcribe ($0.017/minute) or gpt-transcribe ($0.0045/minute). The open-source Whisper model you run yourself is not affected.

Sources: OpenAI API deprecations · OpenAI API pricing

AI tool pricing changes this week (Oct 26, 2026) →