Plan changes
AI plan and pricing changes
Plan, price and feature changes across the products ToolColumn tracks, newest first. Each entry gives the date of the change, the official source and the date it was verified.
November 2026
- Feature
Suno launched Visuals (H3 Move, based on MiniMax H3) on web, iOS, and Android for generating images and music-synced video clips up to 15 seconds, alongside the $120/month Platinum plan ($100/month billed $1,200/year) with 15,000 song credits, 20,000 visual credits, and 120 song downloads a month, plus standalone non-expiring Visual credit packs.
- Feature
Speechify API Simba 3.5 (`simba-3.5`) now accepts the `pronunciations` field on English speech and streaming requests, validating IPA phones against a 110-phone English set (returning HTTP 400 validation_failed unbilled on invalid phones) and listing `en` in `pronunciation_languages` on GET /v1/audio/models.
- Feature
OpenCode v1.19.6 adds an opt-in background upgrade prompt from OpenCode 1 to OpenCode 2.0 preserving sessions, projects, and settings (OpenCode 1 plugins need updates), replaces oldest session images past 25 MiB with placeholders, and auto-compacts oversized request bodies.
- Feature
Kling AI released Kling Image 4.0 for single-image generation and cohesive multi-image series.
- Feature
Devin Desktop v3.14.30819 launched Devin Bots for outcome-driven multi-session work, added tabbed Agent Command Center windows with independent history, moved Devin Local sessions to a background process shared with the CLI, added PDF, CSV, TSV, and media previews in the Files tab, enabled IDE browser-preview element and console picking, retained images in Devin Local up to the provider request limit, and added enterprise proxy auto-detection.
- Feature
Claude Desktop now supports choosing an update release channel: Stable remains the default with slower, more-tested releases, while Latest delivers new versions about once a day.
- Update
Availability: Astra for Law ranked first in a launch evaluation across 24 areas of law, and every ChatGPT Enterprise customer can now request access.
- Feature
Codex added GPT-6.2 Astra (200 input, 10 cached and 1,000 output credits per 1M tokens) and GPT-6.2 Sol (50 input, 2.5 cached and 250 output credits) across Plus, Pro, Business, Enterprise and Edu. Plus and Standard Business five-hour estimates rise to 6–50 messages for GPT-6.2 Astra and 17–180 for GPT-6.2 Sol, and Pro 200 gains Ultrafast on GPT-6.2 Sol (6x included usage; 3x credits through February 28, 2027).
- Feature
OpenAI launched GPT-6.2 Astra and GPT-6.2 Sol in ChatGPT Work, Codex and the API on Plus, Pro, Business, Enterprise and Edu (off by default in Enterprise and Edu; not in regular ChatGPT Chat). Pro 200 now includes Ultrafast starting with GPT-6.2 Sol, while GPT-6.2 Astra cuts standard API prices 20% to $8/$40 per million tokens ($0.40 cached) and GPT-6.2 Sol stays $2/$10 ($0.10 cached).
- Feature
Kiro CLI 3.4.0 adds `PreToolUse` `updatedInput` hooks and `PermissionRequest` `allow`/`deny` hooks in V3 sessions, four `/memories` choices, expanded cloud session resume/listing, and fixes for headless, context window, MCP, search, `/tangent`, and `/rewind` (alongside Enterprise OpenTelemetry real-time events and PrivateLink/VPC Lattice delivery).
- Feature
Kiro IDE 1.4.215 adds conversation forking from any turn, per-run workflow credit reporting, clearer Memory level names, concurrent chats for custom agents or Plan mode, and cloud sessions from the IDE for Europe (Frankfurt) profiles (following Europe Frankfurt Kiro Web launch on November 13, 2026).
- Feature
Kiro Web now includes an interactive workspace terminal beside or below the conversation to run builds, tests, and shell commands.
- Feature
Hermes Agent v0.23.2 rolled up ~1,989 PRs across CLI, installers, Docker, and Hermes Cloud, adding durable subagents across gateway restarts, a per-session prompt-cache clock, `hermes fallback check`, `hermes update --force`, Telegram/Discord one-time browser_vault codes, shared Telegram DM threads, default Gemini Interactions API for Gemini 3+ and Gemini 4 Carbon, fast mode for Claude Mythos 5.2, per-job cron options, new plugin APIs, and a slimmer macOS bundle.
- Improvement
Starting November 26, 2026, the ElevenAgents API (/v1/convai) enforces server-side request timeouts and may return HTTP 504 during temporary latency spikes.
- Feature
OpenAI introduced computer-use agents, improved clinical answers, and new healthcare plugins for ChatGPT for Healthcare to work across healthcare software.
- Pricing change
OpenAI expanded its ChatGPT for Academic Researchers program to offer 200,000 researchers worldwide free access to its frontier models.
- Feature
BytePlus ModelArk added Dreamina Seedance 2.5 Flash (dreamina-seedance-2-5-flash-261115) at $7.00 per 1M tokens for 480p and 720p ($4.30 with video input; about $0.34 and $0.76 per 5-second 16:9 clip), and added 4K output on Seedance 2.5 at $4.70 per 1M tokens ($2.90 with video input).
- Feature
OpenAI began rolling out ChatGPT Wallet in dot conversations for US ChatGPT Pro subscribers aged 18 and older with eligible US-issued American Express, Chase, Mastercard and Visa credit cards. Each purchase generates a payment token and requires biometric verification; Wallet is not available in Chat, Work or standalone Codex threads.
- Update
Retirement: GitHub will deprecate Claude Fable 5, Claude Haiku 4.5, Gemini 3.8 Flash, Gemini 4 Argon, Grok 4.6 and Kimi K3 across all GitHub Copilot experiences on December 14, 2026. Claude Haiku 4.5 stays available to individual subscribers on annual plans.
- Feature
ChatGPT Wallet lets agents in ChatGPT Work, Codex and dots complete purchases after the user approves them, with the card's existing protections.
- Feature
Rask AI released Rask API v3 and the Rask MCP Server.
- Feature
Recraft Studio added the new MiniMax H3 models: H3 Image, H3 Highspeed and H3 Lite. Recraft published notes on their limits from test runs of six image prompts and four video prompts.
- Feature
Recraft V4.1 is available on Leonardo.Ai, where users can set up to 5 colors and a background color for the images it makes.
- Update
Availability: Kiro Mobile is on Google Play for Android. It starts or continues Kiro cloud sessions from a phone, follows replies and tool activity live and shows each edited file as a diff card; review and merge stay in the Git provider. It requires a Pro, Pro+, Pro Max or Power plan.
- Feature
Google released Gemini 4 Hydrogen, a fast model for everyday coding and agentic work, rolling out in the Gemini app, Google Search, Google AI Studio, the Gemini API and Gemini for business. In the API it costs $1.50 per 1M input tokens and $7.50 per 1M output tokens across its 1M-token context window, with low, medium, high (default) and xhigh thinking levels. Free Gemini app users get it through automatic model selection.
- Feature
Claude Meetings is rolling out in beta on Team and Enterprise plans in Claude and Claude Desktop. Users can add Claude to a Google Meet call as a visible participant that takes notes, answers questions and builds artifacts such as docs, whiteboards and slides during the call.
- Feature
Gemini 4 Hydrogen is available in Google Antigravity to paid Google AI Pro and Ultra subscribers. Google also added Orchestrator to the model selector: Gemini 4 Carbon plans and reviews while Gemini 4 Hydrogen carries out the work in the same conversation, at a fixed configuration with no thinking-level choice.
- Feature
Recraft Studio's Vectorizer Pro, available on every plan since October 29, turns PNG, JPG and WebP images into cleaner SVGs than the standard Vectorizer: straighter lines, sharper corners and smoother curves.
- Feature
Anthropic launched Claude Whiteboards and Claude Sheets in beta and made Claude Dashboards generally available as Claude Data across all paid plans on web, desktop and mobile. On Team plans all three are on by default; on Enterprise plans Whiteboards and Sheets are off by default, and Claude Data turns on by default on November 20, 2026.
- Feature
Gemini 4 Carbon became generally available in the Gemini API as gemini-4-carbon, on the paid tier only, at an introductory $2 per 1M input tokens and $10 per 1M output tokens through November 30, 2026 and $4/$20 from December 1. In the Gemini app it is listed on Google AI Ultra.
- Feature
Kiro credit pooling now works for enterprise profiles that connect directly to Okta or Microsoft Entra ID, in addition to AWS IAM Identity Center. Administrators add users and groups under Access management > Credit sharing, and new profiles on any of the three identity sources start with credit pooling.
- Feature
Kiro Enterprise Management entered preview: administrators write one policy in the AWS console that Kiro applies in the IDE, CLI, Kiro Web and the new Kiro Crew LTS edition, with rules for models, shell commands, files, web search and MCP tools. It is open only to Kiro profiles newly created in US East (N. Virginia).
- Feature
Grok 4.7 is available in Kiro with experimental support in the IDE, CLI, Crew and Web for Pro, Pro+, Pro Max and Power subscribers. Requests up to 200K tokens use credits at 2.7x and longer requests at 5.4x, within a 500K-token context window.
- Pricing change
invideo AI's Max plan now costs $150 a seat a month, or $75 billed annually, with 5,000 credits a month. It was $250 for 8,000 credits with no annual discount. The 5,000-credit Plus 2x tier at $150 is no longer listed, and Max 2x and Max 3x keep their prices.
- Feature
GPT Image 2.6 Flare and GPT Image 2.6 Sunburst are available on Leonardo.Ai for refining one image edit by edit, with up to 16 image references.
- Feature
Scribe v3 and Scribe v3 Live are available in the Speech to Text API, for recorded and real-time transcription. They succeed Scribe v2 and Scribe v2 Realtime, which remain available. Scribe v3 adds number formatting and secondary language codes; Scribe v3 Live adds turn prediction, a multilingual mode and a keyterm library of up to 10,000 terms.
- Feature
Claude Code supports Claude Fable 5.2 from version 2.1.321: the fable alias now resolves to Fable 5.2 (to Fable 5 in Claude apps gateway sessions), and Fable 5.1 stays selectable by its model ID. No Fable model is the default on any plan or provider. Fable 5.2 also offers a fast mode in Claude Code, billed at $16/$80 per 1M tokens.
- Feature
Anthropic released Claude Fable 5.2 and Claude Mythos 5.2, the same underlying model, at $8 per 1M input tokens and $40 per 1M output tokens, down from $10 and $50 for Fable 5.1 and Mythos 5.1, with cache hits at $0.20 and an optional fast mode at $16/$80. Fable 5.2 is available on every paid Claude plan under the existing Fable rules (up to 50% of weekly usage on Max, Ultra and Team Premium seats; usage credits on Pro and Team Standard seats) and on the Claude API, Amazon Bedrock, Google Cloud and Microsoft Foundry as claude-fable-5-2. Mythos 5.2 is limited to organizations verified through Anthropic's Cyber and Life Sciences Verification Programs. Fable 5.1 remains available as a legacy model.
- Feature
OpenAI released GPT Image 2.6 Sunburst and GPT Image 2.6 Flare in the API for image generation and editing, at the same token rates as GPT Image 2.5: $5 per 1M text input tokens, $8 per 1M image input tokens and $30 per 1M image output tokens. Both add an adjustable complexity setting, SVG output and animated GIF output, and markup-guided edits now work on GPT Image 2.5 and 2.6. OpenAI names the 2.6 models as replacements for gpt-image-1-mini, gpt-image-1.5 and chatgpt-image-latest, which shut down on December 1, 2026; GPT Image 2.5 and GPT Image 2 remain available.
- Feature
Perplexity launched Teammates, personal agents built on Computer that own ongoing work: each remembers project goals, task history and shared files across runs, can pass work to other Teammates, and can join a Slack channel where colleagues hand it work or ask for updates. Teammates launched on the web for Max subscribers, with iOS to follow; Free and Pro subscribers have to upgrade to Max to use them.
- Pricing change
VEED raised its yearly per-user prices to $12 a month for Creator and $22 for Pro, while monthly billing fell to $25 and $45. Pro now includes 30,000 AI credits a year, up from 18,000, and a new Studio plan costs $39 a month billed yearly or $79 monthly, with 180,000 credits a year.
- Update
Retirement: Google moved the removal of Claude Sonnet 4.6, Claude Opus 4.6 and GPT-OSS-120b from Antigravity to November 18, 2026, from the November 2 date it first gave. Until then they stay on the free tier, Google AI Plus and Google AI Pro, and GPT-OSS-120b also on Ultra.
- Feature
SpaceXAI released Grok Bot 1.0. Bots can now be shared with a team and brought into Slack, Microsoft Teams and video calls, offer to take on work they recognize, and handle browser tasks, documents and presentations more reliably. Team Bots are on Team and Enterprise plans; Meetings are on Team, Enterprise and Ultra.
- Pricing change
ElevenLabs now bills Architect usage from workspace credits as its own line item, separate from agents' minutes and LLM cost. Billing is rolling out gradually, and the dashboard composer shows when it applies to a workspace.
- Feature
Professional Voice Clone fine-tuning for Eleven v4 and Eleven v4 Turbo is available to every workspace. Eleven v4 requests for a fine-tuned clone now use the clone strength saved in the voice's settings.
- Feature
SFX v3, ElevenLabs' newest sound effects model, is available to every workspace and is now the default model in Sound Effects, Studio and Flows. It follows prompts with several instructions more closely and handles music or vocals better than SFX v2, which will be removed on March 1, 2027.
- Update
Retirement: xAI retired the grok-imagine-image-quality and grok-imagine-image-pro API models. Requests to them are now served by grok-imagine-image-2.0 at low quality, at a lower per-image price, with no request changes needed.
- Update
Availability: Synthesia now works inside ChatGPT and Claude. From a prompt, an attached deck or document, or connected tools, users can create an on-brand video in the chat, edit it by asking, and share a link or download the file. Videos are made with Syren by default.
- Feature
Murf added two Speech to Text models to its API. Finch-Indic transcribes English (India) and 11 Indian languages; Finch-Global transcribes English (US, UK and Australia), German, Spanish, French, Italian, Portuguese and Dutch. Speech to Text requests now have to name the model on the WebSocket URL.
October 2026
- Feature
Claude can now create a design, deck or document in any conversation, including in Claude Code and the Artifacts tab. Artifacts, including Claude Design, Claude Slides and the new Claude Docs, are available on every plan including Free; on Enterprise plans Design, Slides and Docs are in beta.
- Update
Retirement: Fast mode for Opus 4.8 is deprecated as of October 30, 2026 and will be removed on November 30, 2026; after removal such requests return an error. Users are directed to Opus 5.5.
- Update
Anthropic renamed its Max plans: Max 5x is now Max ($100 a month) and Max 20x is now Ultra ($200 a month), with no price change and no action needed from subscribers. Ultra drops the five-hour limit for verified accounts and gives 10 times Pro's usage every week; Anthropic says weekly limits are unchanged, and newer accounts keep a temporary limit until they are verified.
- Pricing change
ChatGPT Pro tiers now include $100, $200 or $500 of monthly OpenAI API credits, and Business includes $20 per paid Standard seat and $100 per Premium seat, capped at $500 per workspace, after an owner links an API organization.
- Pricing change
GitHub Enterprise customers can now use all Copilot features without a separate Copilot subscription and pay for the AI credits they use. This Copilot access includes no monthly AI credits per seat, so usage is metered from the first request against the enterprise budget; it is sold through GitHub Sales for now, with self-serve dates to follow. Outside GitHub Enterprise, Copilot Standalone costs $10 per user per month.
- Feature
GitHub made Copilot Fusion generally available in VS Code, Copilot CLI, the GitHub Copilot app and JetBrains IDEs: a router that picks the models and execution pattern for each task and returns one response. Fusion Standard is on every Copilot plan, while Fusion High and Max need Pro+, Max, Business or Enterprise. Fusion bills each model it uses at that model's standard rate in AI credits, with no separate fee, and paid plans get 30% off those model costs through November 30, 2026.
- Feature
OpenAI started rolling out texting for dots in beta to ChatGPT Pro users in the US, over iMessage on iPhone or RCS on Android. A dot can text first, and the chats carry text, photos and voice messages; OpenAI sends them through Linq, a third-party messaging provider. To turn it on, open the dot's profile in ChatGPT and select Texting.
- Feature
Smart reports, in-app surveys and pooled group budgets are now generally available on Claude Enterprise. Pooled group budgets let admins set one shared budget per team.
- Feature
Google announced Gemini 4 Carbon, its new frontier model. It is not in the Gemini app yet; access starts this week with a subset of cyber defenders and widens to trusted partners through Google's Fairwind Program.
- Feature
VEED launched its Dubbing API, which translates a video into another language in each speaker's cloned voice and lip-syncs the footage. Standard costs $0.11 per second of input video and dubs into 100 languages; Pro costs $0.18 per second, dubs into 94 languages and adds 14 accent variants, kept background audio and multi-speaker videos. Videos can run up to 10 minutes, billing is usage-based with no seat fees or minimums, and the API runs on the VEED API or through fal.
- Feature
Muse became available in Singapore on iOS, Android and muse.ai, free for most of what people need, with AI glasses support coming soon. Connectors for partners including Shopee, Grab, Singapore Airlines, Agoda, ShopBack, Klook, Trip.com and Lazada roll out progressively.
- Feature
Runway added Runway Ads, included with every paid plan. It reads a user's best-performing ads on Meta, Google and TikTok, generates on-brand variants for each placement and publishes the ones the user approves.
- Pricing change
Text to Speech and Speech to Text requests priced below one credit are now charged one credit.
- Pricing change
Business plans on current pricing now pay the same image and video credit rate as other self-serve plans, about 17% fewer credits per image and per video second than before.
- Feature
GPT-6.1 Terra is now available in Codex on Plus, Pro, Business, Enterprise and Edu. It uses 25 credits per million input tokens, 1.25 cached and 125 output, half of GPT-6.1 Sol's rate, and OpenAI estimates 35–285 local messages per five hours on Plus and Standard Business seats against 15–160 for GPT-6.1 Sol. Overall Work and Codex plan limits are unchanged.
- Feature
OpenAI released GPT-6.1 Terra, a fast, efficient model for visual reasoning, precise computer use and clear writing. In ChatGPT it is available in ChatGPT Work and Codex on Plus, Pro, Business, Enterprise and Edu, chosen under Advanced in the model list, and not in Chat; Enterprise access depends on workspace availability and administrator controls. In the API, gpt-6.1-terra costs $1 per million input tokens and $5 per million output tokens, half of GPT-6.1 Sol.
- Feature
Meta launched Muse in Japan on iOS, Android and muse.ai, its first region outside North America, with most basic features free and subscription plans for heavier use. Gmail, Google Calendar and Instagram connectors and partners including Stripe, Shopify, Trip.com, Agoda and Klook came at launch; Mercari, Money Forward ME and Ticket Pia integrations are in progress.
- Feature
Runway added App Studio on paid plans. You describe the creative app you need in plain language, and App Studio builds it with Runway's models, your brand and your files, then hosts it.
- Feature
Claude Enterprise plans now get the unified Claude experience that combines chat and Cowork in one conversation on web, desktop and mobile, after Pro and Max got it in September. Agentic work runs in cloud sessions that Anthropic manages, and Enterprise pricing stays usage-based. Admins get a new Efficient mode that keeps the unified experience on Sonnet and Haiku models, alongside model entitlements by role and spend limits at the organization, group and member level.
- Feature
Ideogram launched Ideogram Vector 1.0, its first vector model. It creates editable SVGs from a text prompt, with optional reference images, or converts an existing image into a vector without a prompt, in the Ideogram app, on the API and on partner platforms Ideogram does not name. On the API an SVG from a prompt costs $0.13 at Low quality, $0.16 at Medium and $0.20 at High, or $0.32 at High with reference images; vectorizing an image costs $0.10. Ideogram does not publish which app plans include Vector or its app credit cost.
- Feature
Higgsfield launched Content Factory, which turns a "What if…" idea into a faceless video of 15 seconds to 5 minutes, with one to four episodes per generation. It runs on higgsfield.ai, in Claude through the Higgsfield MCP with /content-factory, and in Supercomputer as a skill. Each episode costs 15.75 credits for 15 seconds, 63 for a minute or 315 for 5 minutes.
- Update
Availability: Anthropic made Claude Models on Google Cloud (Anthropic managed) generally available. Google Cloud customers get the full Claude Platform API, including Claude Managed Agents, code execution and prompt caching, with Google Cloud IAM sign-in and billing on their Google Cloud invoice that draws down existing Google Cloud commitments. Fast Mode on Claude Opus 5.5 is included at $8/$40 per million input/output tokens. It runs on US and Global endpoints, and the Google-managed offering remains for customers who need Google Cloud as the data processor.
- Feature
Murf launched Speech to Text (Beta) in its API: real-time transcription over WebSocket for voice agents, IVR and call analytics in 12 languages (English (India) and 11 Indian languages) with automatic language detection, at $0.24 per hour billed per second. The free tier allows 5 concurrent requests; pay as you go defaults to 20 and goes up to 150.
- Feature
Kling 4.0 is available on Higgsfield for text, image and reference-based video and for editing finished clips. Higgsfield applies a limited-time launch discount and shows the exact credit cost on the Generate button; the pricing page still lists Kling 3.0 rates only.
- Update
Retirement: Lovable retired the Gemini 2.5 Pro, Gemini 2.5 Flash and Gemini 2.5 Flash Lite chat models for the AI features in apps built on Lovable on October 20, 2026. Requests for Gemini 2.5 Pro and Flash now go to Gemini 3.8 Flash, and Flash Lite requests go to Gemini 3.1 Flash Lite. Apps keep working without a code change, but the replacement model can cost a different amount per request and respond differently.
- Feature
Cursor's Agent SDK is generally available. You define an agent as a folder of code, test it with evals, and deploy it to Cursor or your own infrastructure.
- Pricing change
Krea brought back its Basic plan at $9 a month, or $5.25 billed yearly ($63 a year), with 5,000 compute units a month and selected Seedance video models, so paid plans start with Basic again rather than Pro at $35 a month, or $21 billed yearly.
- Feature
ElevenLabs introduced Eleven Realtime, a model family for ElevenAgents voice agents rolling out gradually across workspaces. Each model is a complete agent configuration that covers the LLM, Speech to Text and Text to Speech: v1 Mini is the default and the fastest and cheapest, and v1 Max reasons before it answers. It works with existing OpenAI Realtime API clients. The same release added Realtime Translation for enterprise workspaces and a Video to Music API.
- Pricing change
Murf's Studio Creator plan now includes 24 hours of voice generation a year on annual billing, down from 48, at the same $19 a month ($228 a year). Studio Business keeps 96 hours a year at $66 a month ($792 a year) and now allows 500 projects.
- Pricing change
Kling replaced its Ultra membership (26,000 credits for $159.99 a month) with Ultra+, which includes 57,000 credits a month for $544.99 after a $402 first month, or $4,902.99 a year. Premier now offers larger credit tiers up to 38,000 credits, with 18,000 credits at $181.99 a month; its 8,000-credit tier stays at $80.96. Standard and Pro are unchanged.
- Pricing change
MiniMax added Scale and Business tiers to its audio subscription, at $249 and $999 a month ($2,390 and $9,590 a year), with 3.3 million and 20 million audio points a month. Starter, Standard and Pro keep their prices of $5, $30 and $99 a month but now include 100,000, 300,000 and 1.1 million audio points.
- Pricing change
Krea dropped its Basic plan ($9 a month, or $5.25 billed yearly), so paid plans now start with Pro at $35 a month, or $21 billed yearly. Business is priced by compute units, from $50 a month (or $40 billed yearly) for 20,000 units up to 1.5 million units, with up to 50 seats included.
- Pricing change
VEED dropped its Studio plan and repriced Creator and Pro. Creator costs $26 per user a month, or $10 billed annually, with 6,000 credits and 120 subtitle hours a year. Pro costs $46 per user a month, or $20 billed annually, with 18,000 credits and 4K export. A free tier remains, and Enterprise is custom.
- Pricing change
Synthesia replaced its Creator plan with Pro and cut the price of Starter. Starter costs $19 a month, or $14 a month billed yearly, with 1,250 credits a month (15,000 a year). Pro costs $89 a month, or $59 billed yearly, with 6,000 credits a month (72,000 a year). Basic stays free with 500 credits a month, and Enterprise is custom.
- Update
Availability: Pika launched a Pika x ChatGPT plugin extension that brings its 40+ creative apps and 100+ generative media models for images, video and sound into ChatGPT on desktop and web, with work saved back to the user's Pika library.
- Pricing change
Cursor is now sold through The X Subscription with Grok and X Premium: Free, Plus at $8 (Grok models only), Premium at $30, Super at $100 and Ultra at $200 a month, with one weekly pool shared with Grok. Third-party models are billed at API list price from prepaid credit on Premium and above, and team seats cost $30, $100 or $200 per user. Existing Cursor plans continue until subscribers switch.
- Pricing change
SpaceXAI launched The X Subscription, which sells Grok, Grok Bot, Cursor and X Premium as one plan: Free, Plus at $8, Premium at $30, Super at $100 and Ultra at $200 a month, with one weekly usage pool shared across Grok and Cursor (4x, 15x and 50x Plus). Grok Bot and video generation start at Premium. SuperGrok plans are closed to new sign-ups; existing subscribers can keep them or switch with a prorated refund.
- Feature
Higgsfield released Katana, a skill for Claude that rebuilds a reference video edit with your own characters, products or clothing through the Higgsfield MCP connector. It needs a paid Higgsfield plan and bills plan credits, with no separate MCP fee.
- Feature
OpenAI added Ultrafast for GPT-6.1 Sol. Pro 500 now includes Ultrafast for both GPT-6 Astra and GPT-6.1 Sol, using included usage at 8x the Standard rate and then credits at 6x, and the API offers GPT-6.1 Sol with service_tier "ultrafast" at 6x Standard pricing. Pro 100 and Pro 200 still do not include Ultrafast, and buying credits does not unlock it.
- Feature
Claude Max and Team now include monthly API credits: $100 on Max 5x, $200 on Max 20x, and $20 per Team Standard seat or $100 per Team Premium seat, pooled and capped at $500. They cover the Claude API, Agent SDK and claude -p with a linked organization's API key, not interactive Claude Code, and do not roll over; Free, Pro and Enterprise are not eligible.
- Feature
ChatGPT Chat moved to GPT-6 with Intelligent UI: Plus, Pro, Business and Enterprise use GPT-6 Sol and Free and Go use GPT-6 Luna, with Free and Go rolling out from October 8. The models behind ChatGPT Work and Codex did not change, and Pro reasoning stays on GPT-6 Astra.
- Update
ToolColumn modeling correction: Claude Code's Anthropic API access is now listed as one usage-based route billed at Anthropic API rates instead of separate per-model plans. Per-model rates are listed on Anthropic's pricing page and in the Claude pricing guide. Prices did not change.
- Feature
Claude Code added Claude Haiku 5.5 as the default Haiku model on the Anthropic API, so the haiku alias now selects it, with a 1M-token context window at $0.10/$0.50 per million tokens ($0.50/$2.50 for prompts over 100,000 tokens).
- Feature
Anthropic released Claude Haiku 5.5 for high-volume, latency-sensitive work at $0.10/$0.50 per million tokens for prompts up to 100,000 tokens and $0.50/$2.50 above, with a 1M-token context window, on the Claude API, AWS, Google Cloud and Microsoft Foundry. Haiku 4.5 remains available as a legacy model, and Sonnet 5.5 cache reads were cut from $0.20 to $0.10 per million tokens.
- Feature
Google made Nano Banana 2.1 (gemini-nano-banana-2.1) generally available in the Gemini API as the update to Nano Banana 2. In the same release it deprecated Nano Banana 2 (gemini-3.1-flash-image), with no shutdown date announced, and told developers to migrate to Nano Banana 2.1.
- Feature
Claude Code added mods, plugins made of TypeScript or JavaScript event handlers that change how Claude Code looks and behaves, installed from a marketplace with /plugin install. They need CLI v2.1.287 or later or the Desktop app's Code tab from v2.1.286, and also run in the VS Code extension. Mods aren't sandboxed: they run with your permissions and can spend your plan or API usage.
September 2026
- Feature
OpenAI released GPT-6.1 Sol in Codex, rolling out from Pro to Plus, Business, Enterprise and Edu, at 50 input, 2.5 cached and 250 output credits per million tokens. The refreshed Codex CLI added voice steering and an /agents view, and Codex Cloud added reusable environments for Plus and higher plans.
- Feature
OpenAI introduced dots, always-on agents powered by GPT-6 Astra for Pro and Business Premium; added the $500 Pro 500 plan with 25x Plus usage and GPT-6 Astra Ultrafast; reopened Pro 200 to new subscriptions with a lower included allowance; and began rolling out GPT-6.1 Sol in ChatGPT Work and Codex, starting with Pro.
- Feature
Claude Code added Claude Sonnet 5.5 as the default Sonnet model on the Anthropic API, so the sonnet alias now selects it, with a 1M-token context window at $2/$10 per million tokens.
- Feature
Anthropic released Claude Sonnet 5.5 at the same $2/$10 per million token price as Sonnet 5, with faster output, a 1M-token context window, and availability on the Claude API, AWS, Google Cloud, and Microsoft Foundry. Sonnet 5 remains available as a legacy model.
- Feature
Claude Code 2.1.280 added Claude Opus 5.5 (claude-opus-5-5) and made it the default Opus model.