Learn

Compare AI Companies

Who makes today's major AI models, what each company sells, and how the products line up. Start from a question, compare side by side, or open one company's full profile.

Select any mark to see its note and the company page behind it. Facts checked through November 2, 2026; this field moves monthly, so check a source before you commit to anything.

Start With a Question

Pick the question closest to yours. The comparison below narrows to the rows that answer it.

Pick a coding agentWhich companies offer an agent that works in my code base, and how far can I extend it?
Run models myselfWhose models can I download and run on my own hardware, and under what license?
Build on an APIWhose API fits a product I am building, and what lowers the bill?
Roll out at workWhat does an organization get for data protection, administration, and where it runs?
Choose a chat appWhich assistant apps can research, talk, remember, and make images or video?
Prepare for an interviewWhat should I know about each company before I talk about its products?

Compare Companies

Capability group

AppsBuildAgentsBusiness

What a person can use without writing code.

Offered
Limited
Not offered

Select a cell to see what is limited and where the fact comes from. The detail opens under that company's row; Escape closes it.

Apps capabilities by company
Company	Chat app	Deep research	Live voice	Image generation	Video generation	Memory
Anthropic
OpenAI
Google
Meta
xAI
Microsoft
Amazon
Mistral AI
DeepSeek
Alibaba
Moonshot AI

Chat app:: A web or phone assistant anyone can sign up for.
Deep research:: Searches many sources and writes a cited report.
Live voice:: Spoken conversation in real time.
Image generation:: Makes images from a description.
Video generation:: Makes video clips from a description.
Memory:: Remembers facts across separate chats.

Compare Models

Each company's current models, plus earlier ones it still offers. Prices are list prices in US dollars per million tokens at the standard tier and shortest context band. Batch, caching, and long-prompt prices differ, so open the company's pricing page before you budget. A dash means no per-token price: the model is priced per image, second, or minute, or is only downloadable. Sort with the column headings, or the Sort menu on smaller screens: newest, largest context, or lowest price first.

Company

Sort

78 of 78 models, newest first

Models, newest first
Model	Company					Weights	About
GPT-6.1 TerraOpenAI, Oct 2026	OpenAI	Oct 2026	$1.00	$5.00	1.05M	Hosted only	A fast, efficient model for visual reasoning, computer use, and writing.`gpt-6.1-terra`
DeepSeek-V4.1-Flash-1017DeepSeek, Oct 2026	DeepSeek	Oct 2026	$0.30	$1.20	1M	MIT	Flash model served in the app and API, with thinking and non-thinking modes and image input. The 1017 update replaced the September release.`deepseek-flash`
Muse Spark OpenMeta, Oct 2026	Meta	Oct 2026	No per-token price	No per-token price	1M	Muse Model License	Open-weight Muse Spark (816B total, 42B active) for self-hosting.
Muse WonderPreviewMeta, Oct 2026	Meta	Oct 2026	$4.00	$20.00	1M	Hosted only	Most capable API model, for long agent tasks and software work.`muse-wonder-1.0`
Claude Haiku 5.5Anthropic, Oct 2026	Anthropic	Oct 2026	$0.10	$0.50	1M	Hosted only	Fastest model, for high-volume, latency-sensitive tasks such as classification and routing.`claude-haiku-5-5`
Nova 2.5 SonicAmazon, Oct 2026	Amazon	Oct 2026	No per-token price	No per-token price	256K	Hosted only	Speech-to-speech model for real-time voice agents that call tools.`amazon.nova-2-5-sonic`
Gemini 4 ArgonGoogle, Oct 2026	Google	Oct 2026	$2.00	$10.00	1M	Hosted only	Flagship reasoning model for complex coding, math, science, and multi-step agent tasks. The listed price rises to $4 and $20 on December 1, 2026.`gemini-4-argon`
Nano Banana 2.1Google, Oct 2026	Google	Oct 2026	No per-token price	No per-token price	128K	Hosted only	High-efficiency image generation and conversational image editing.`gemini-nano-banana-2.1`
MAI-Code-1.1-FlashPreviewMicrosoft, Oct 2026	Microsoft	Oct 2026	No per-token price	No per-token price	256K	Hosted only	Agentic coding model in GitHub Copilot. Microsoft says it can be downloaded and run locally but lists no license or download location.`MAI-Code-1.1-Flash`
MAI-Voice-2.1PreviewMicrosoft, Oct 2026	Microsoft	Oct 2026	No per-token price	No per-token price	Not stated	Hosted only	Speech from text in 23 languages, with emotion styles and gated voice cloning from a short clip.`MAI-Voice-2.1`
Mistral Large 4PreviewMistral AI, Oct 2026	Mistral AI	Oct 2026	$0.70	$2.10	1M	Mistral Ecosystem Open Weight License	General-purpose multimodal model built on a mixture-of-experts architecture.`mistral-large-4`
Voxtral Small Transcribe 3PreviewMistral AI, Oct 2026	Mistral AI	Oct 2026	No per-token price	No per-token price	32K	Hosted only	Transcription and speaker labeling; priced per minute of audio.`voxtral-small-transcribe-3-0`
Voxtral TTS 3PreviewMistral AI, Oct 2026	Mistral AI	Oct 2026	No per-token price	No per-token price	Not stated	Hosted only	Streaming text to speech with voice cloning in more than 35 languages; priced per character.`voxtral-mini-tts-3-0`
Kimi K3.1 ProMoonshot AI, Oct 2026	Moonshot AI	Oct 2026	$2.00	$10.00	1M	Hosted only	Flagship for long coding and knowledge work, post-trained further from K3; weights promised by 4 November 2026. A Highspeed version runs 2 to 3 times faster at twice the price.`kimi-k3.1-pro`
GPT-6.1 SolOpenAI, Sep 2026	OpenAI	Sep 2026	$2.00	$10.00	1.05M	Hosted only	Close to Astra on complex work and coding, at a lower price.`gpt-6.1-sol`
Claude Sonnet 5.5Anthropic, Sep 2026	Anthropic	Sep 2026	$2.00	$10.00	1M	Hosted only	Combines speed and capability.`claude-sonnet-5-5`
Claude Opus 5.5Anthropic, Sep 2026	Anthropic	Sep 2026	$4.00	$20.00	1M	Hosted only	Suggested starting model for most work, built for long-running agentic coding and knowledge work.`claude-opus-5-5`
GPT-6 LunaOpenAI, Sep 2026	OpenAI	Sep 2026	$0.10	$0.50	1.05M	Hosted only	Most efficient model, for focused, high-volume tasks.`gpt-6-luna`
Grok 4.7xAI, Sep 2026	xAI	Sep 2026	$2.00	$6.00	500K	Hosted only	Flagship model for coding, agent tasks, and knowledge work.`grok-4.7`
Qwen3.8-Omni-FlashAlibaba, Sep 2026	Alibaba	Sep 2026	$0.15	$0.47	1M	Hosted only	Hosted omni model that understands text, images, audio, and long video; a realtime variant talks back.`qwen3.8-omni-flash`
GPT Image 2.5 SunburstOpenAI, Sep 2026	OpenAI	Sep 2026	No per-token price	No per-token price	Not stated	Hosted only	Image generation and editing; the most capable image model.`gpt-image-2.5-sunburst`
GPT-6 AstraOpenAI, Sep 2026	OpenAI	Sep 2026	$10.00	$50.00	1.05M	Hosted only	Most capable model, for the most demanding work.`gpt-6-astra`
Muse Spark 1.3Meta, Sep 2026	Meta	Sep 2026	$1.25	$4.25	1M	Hosted only	Recommended model for new work, tuned for agent workflows and coding.`muse-spark-1.3`
Claude Fable 5.1Anthropic, Sep 2026	Anthropic	Sep 2026	$10.00	$50.00	1M	Hosted only	For demanding reasoning and long-horizon agent work.`claude-fable-5-1`
Gemini 3.8 FlashGoogle, Sep 2026	Google	Sep 2026	$0.75	$3.75	1M	Hosted only	Newest Flash model, for long-horizon software work, agents, and enterprise workflows. The listed price doubles on January 1, 2027.`gemini-3.8-flash`
MAI-Image-2.6PreviewMicrosoft, Sep 2026	Microsoft	Sep 2026	No per-token price	No per-token price	32K	Hosted only	Makes and edits images from text and up to 5 reference images.`MAI-Image-2.6`
MAI-Transcribe-2PreviewMicrosoft, Sep 2026	Microsoft	Sep 2026	No per-token price	No per-token price	Not stated	Hosted only	Speech recognition across 60 languages, including mixed-language speech.`MAI-Transcribe-2`
Qwen3.8-FlashAlibaba, Aug 2026	Alibaba	Aug 2026	$0.15	$0.47	1M	Hosted only	Hosted fast, low-cost model for high-volume work.`qwen3.8-flash`
Qwen3.8-27BAlibaba, Aug 2026	Alibaba	Aug 2026	$0.50	$3.00	1M	Apache 2.0	Compact dense 27B vision-language model for coding and agent work.`qwen3.8-27b`
DeepSeek-V4-ProDeepSeek, Aug 2026	DeepSeek	Aug 2026	$1.32	$3.96	1M	MIT	Larger text model aimed at agent work and production use. The API serves the 0813 version.`deepseek-v4-pro`
Grok 4.6PreviousxAI, Aug 2026	xAI	Aug 2026	$2.00	$6.00	500K	Hosted only	Prior flagship, tuned for long-running agents and visual, interactive work.`grok-4.6`
Qwen3.8-2.4T-A95BAlibaba, Aug 2026	Alibaba	Aug 2026	$2.00	$6.00	1M	Qwen3.8-Max License (custom)	Open checkpoint of the Max-class model: a 2.4T mixture-of-experts text model with thinking always on.`qwen3.8-2.4t-a95b`
Wan3.0-VideoAlibaba, Aug 2026	Alibaba	Aug 2026	No per-token price	No per-token price	Not stated	Hosted only	Hosted video generation model, priced per second of output rather than per token.`wan3.0-video`
Muse Spark 1.2PreviousMeta, Aug 2026	Meta	Aug 2026	$1.25	$4.25	1M	Hosted only	Previous Spark version and the default in Muse Code; suggested for audio input.`muse-spark-1.2`
Qwen3.8-MaxAlibaba, Aug 2026	Alibaba	Aug 2026	$2.00	$6.00	1M	Hosted only	Hosted flagship for complex reasoning, coding, and long agent tasks.`qwen3.8-max`
Gemini Omni 1.1 FlashGoogle, Aug 2026	Google	Aug 2026	No per-token price	No per-token price	1M	Hosted only	Makes and edits short video clips from text, images, and video through conversation.`gemini-omni-1.1-flash`
Muse GlimmerMeta, Aug 2026	Meta	Aug 2026	No per-token price	No per-token price	128K	Apache 2.0	Small 30B open-weight model distilled from Muse Spark, for local and edge use.
Grok Imagine Image 2.0xAI, Aug 2026	xAI	Aug 2026	No per-token price	No per-token price	Not stated	Hosted only	Image generation and editing, priced per image.`grok-imagine-image-2.0`
MAI-Thinking-1PreviewMicrosoft, Aug 2026	Microsoft	Aug 2026	No per-token price	No per-token price	256K	Hosted only	Reasoning model for multi-step work, coding, and long documents.`MAI-Thinking-1`
Claude Opus 5PreviousAnthropic, Jul 2026	Anthropic	Jul 2026	$5.00	$25.00	1M	Hosted only	Earlier Opus model, still offered as a legacy option.`claude-opus-5`
Muse ImageMeta, Jul 2026	Meta	Jul 2026	No per-token price	No per-token price	Not stated	Hosted only	Makes and edits images from text and reference images; priced per image.`muse-image-1.0`
GPT-Realtime-2.1OpenAI, Jul 2026	OpenAI	Jul 2026	$4.00	$24.00	128K	Hosted only	Speech in and speech out in real time, with reasoning and tool use.`gpt-realtime-2.1`
Gemini 3.5 Flash-LiteGoogle, Jul 2026	Google	Jul 2026	$0.30	$2.50	1M	Hosted only	Low-cost, low-latency model for high-volume work and document parsing.`gemini-3.5-flash-lite`
Gemini 3.6 FlashPreviousGoogle, Jul 2026	Google	Jul 2026	$0.75	$3.75	1M	Hosted only	Previous-generation Flash model for agentic coding and execution. The listed price doubles on January 1, 2027.`gemini-3.6-flash`
OCR 4.1Mistral AI, Jul 2026	Mistral AI	Jul 2026	No per-token price	No per-token price	Not stated	Hosted only	Document OCR with paragraph bounding boxes, block labels, and confidence scores; priced per page.`mistral-ocr-4-1`
Kimi K3PreviousMoonshot AI, Jul 2026	Moonshot AI	Jul 2026	$3.00	$15.00	1M	Kimi K3 License (custom; extra terms for large commercial use)	Previous flagship with 2.8 trillion parameters, for long coding, knowledge work, and reasoning.`kimi-k3`
Claude Sonnet 5PreviousAnthropic, Jun 2026	Anthropic	Jun 2026	$2.00	$10.00	1M	Hosted only	Earlier Sonnet model, still offered as a legacy option.`claude-sonnet-5`
Claude Fable 5PreviousAnthropic, Jun 2026	Anthropic	Jun 2026	$10.00	$50.00	1M	Hosted only	Earlier Fable model, still offered as a legacy option.`claude-fable-5`
Qwen3.7-PlusAlibaba, Jun 2026	Alibaba	Jun 2026	$0.40	$1.60	1M	Hosted only	Hosted middle tier that balances quality, speed, and cost; image and video input.`qwen3.7-plus`
Grok Imagine Video 1.5xAI, Jun 2026	xAI	Jun 2026	No per-token price	No per-token price	Not stated	Hosted only	Video with sound from text, an image, or references, priced per second.`grok-imagine-video-1.5`
Kimi K2.7 CodePreviousMoonshot AI, Jun 2026	Moonshot AI	Jun 2026	$0.95	$4.00	256K	Modified MIT	Coding model built on K2.6 that always thinks; a Highspeed version doubles the price for faster output. Closed to new API users; retiring by end of 2026.`kimi-k2.7-code`
Grok Build 0.1xAI, May 2026	xAI	May 2026	$1.00	$2.00	256K	Hosted only	Coding model trained for agent coding work; in early access.`grok-build-0.1`
Fara1.5-9BMicrosoft, May 2026	Microsoft	May 2026	No per-token price	No per-token price	256K	MIT	Small open agent model, built on Qwen3.5-9B, that operates a web browser from screenshots.`microsoft/Fara1.5-9B`
DeepSeek-V4-FlashPreviousDeepSeek, Apr 2026	DeepSeek	Apr 2026	No per-token price	No per-token price	1M	MIT	First V4 small model. Retired from the API in September 2026; the weights remain downloadable.
Gemma 4 31BGoogle, Apr 2026	Google	Apr 2026	No per-token price	No per-token price	256K	Apache 2.0	Largest dense model in the Gemma 4 open family, which runs from phones to servers.`gemma-4-31b-it`
Grok 4.3xAI, Apr 2026	xAI	Apr 2026	$1.25	$2.50	1M	Hosted only	Lower-cost general model with a 1 million token context and adjustable reasoning.`grok-4.3`
Kimi K2.6PreviousMoonshot AI, Apr 2026	Moonshot AI	Apr 2026	$0.95	$4.00	256K	Modified MIT	General model for chat, coding, and agent tasks, with thinking and non-thinking modes. Closed to new API users; retiring by end of 2026.`kimi-k2.6`
Grok 4.20PreviousxAI, Mar 2026	xAI	Mar 2026	$1.25	$2.50	1M	Hosted only	Fast tool-calling model; a multi-agent variant runs several research agents at once.`grok-4.20-0309-reasoning`
Phi-4-reasoning-vision-15BMicrosoft, Mar 2026	Microsoft	Mar 2026	No per-token price	No per-token price	16K	MIT	Small open model that reasons over images, for example math from pictures and reading screens.`microsoft/Phi-4-reasoning-vision-15B`
Mistral Small 4Mistral AI, Mar 2026	Mistral AI	Mar 2026	$0.15	$0.60	256K	Apache 2.0	Hybrid model that combines instruction following, reasoning, and coding in one model.`mistral-small-2603`
Gemini 3.1 Pro (preview)PreviewGoogle, Feb 2026	Google	Feb 2026	$2.00	$12.00	1M	Hosted only	Earlier Pro model in preview, aimed at coding and multi-step agent work.`gemini-3.1-pro-preview`
Qwen3-Coder-NextPreviousAlibaba, Feb 2026	Alibaba	Feb 2026	No per-token price	No per-token price	256K	Apache 2.0	Open 80B mixture-of-experts model (3B active) built for coding agents and local development.`qwen3-coder-next`
Nova 2 LiteAmazon, Dec 2025	Amazon	Dec 2025	$0.33	$2.75	1M	Hosted only	Low-cost multimodal model with optional reasoning, for automation, documents, and support.`amazon.nova-2-lite-v1:0`
DeepSeek-V3.2PreviousDeepSeek, Dec 2025	DeepSeek	Dec 2025	No per-token price	No per-token price	Not stated	MIT	Earlier reasoning model built for agent tool use. Still hosted by third parties such as Amazon Bedrock.
Nova 2 ProPreviewAmazon, Dec 2025	Amazon	Dec 2025	$1.38	$11.00	Not stated	Hosted only	Most capable Nova 2 model for multistep tasks; in preview, with early access for Nova Forge customers.
Ministral 3 14BMistral AI, Dec 2025	Mistral AI	Dec 2025	$0.20	$0.20	256K	Apache 2.0	Largest Ministral 3 model, with text and vision, that can run on local hardware.`ministral-14b-2512`
Ministral 3 3BMistral AI, Dec 2025	Mistral AI	Dec 2025	$0.10	$0.10	256K	Apache 2.0	Smallest Ministral 3 model, with text and vision, for edge deployment.`ministral-3b-2512`
gpt-oss-120bOpenAI, Aug 2025	OpenAI	Aug 2025	No per-token price	No per-token price	128K	Apache 2.0	Open-weight reasoning model that fits on one 80 GB GPU.`gpt-oss-120b`
gpt-oss-20bOpenAI, Aug 2025	OpenAI	Aug 2025	No per-token price	No per-token price	128K	Apache 2.0	Smaller open-weight reasoning model that runs in 16 GB of memory.`gpt-oss-20b`
Grok 2PreviousxAI, Aug 2025	xAI	Aug 2025	No per-token price	No per-token price	Not stated	Grok 2 Community License Agreement	Older 2024 model whose weights are published for local use.
CodestralMistral AI, Jul 2025	Mistral AI	Jul 2025	$0.30	$0.90	128K	Hosted only	Low-latency code completion, including fill-in-the-middle.`codestral-2508`
Llama 4 MaverickPreviousMeta, Apr 2025	Meta	Apr 2025	No per-token price	No per-token price	1M	Llama 4 Community License	Earlier open-weight mixture-of-experts model (400B total, 17B active).
Llama 4 ScoutPreviousMeta, Apr 2025	Meta	Apr 2025	No per-token price	No per-token price	10M	Llama 4 Community License	Smaller Llama 4 open-weight model (109B total, 17B active) with a long context.
Phi-4-mini-instructPreviousMicrosoft, Feb 2025	Microsoft	Feb 2025	No per-token price	No per-token price	128K	MIT	Small text model for low-memory and low-latency use; also offered on Foundry.`microsoft/Phi-4-mini-instruct`
DeepSeek-R1PreviousDeepSeek, Jan 2025	DeepSeek	Jan 2025	No per-token price	No per-token price	128K	MIT	First-generation reasoning model, released with 6 smaller distilled models.
Nova ProPreviousAmazon, Dec 2024	Amazon	Dec 2024	$0.80	$3.20	300K	Hosted only	First-generation multimodal model balancing accuracy, speed, and cost.`amazon.nova-pro-v1:0`
Nova LitePreviousAmazon, Dec 2024	Amazon	Dec 2024	$0.060	$0.24	300K	Hosted only	First-generation low-cost multimodal model for image, video, and text input.`amazon.nova-lite-v1:0`
Nova MicroPreviousAmazon, Dec 2024	Amazon	Dec 2024	$0.035	$0.14	128K	Hosted only	Text-only model for fast, cheap tasks such as summaries, translation, and classification.`amazon.nova-micro-v1:0`

Company Profiles

Anthropic

AI lab, San Francisco, California

Anthropic makes the Claude family of models and offers them through the Claude apps, the Claude Code coding agent, and a developer API. It positions its models for long-running agentic coding and knowledge work, and also offers them through AWS, Google Cloud, and Microsoft Foundry.

Known for Claude, Claude Code, Fable, Opus, Sonnet, Haiku. Checked against the sources below on November 2, 2026. Company site

Products

Apps: Claude · web, desktop, phone
The general assistant app for chat, research, files, and connected tools. Cowork task features now run inside it.
Coding: Claude Code · terminal, code editor, desktop, web, phone
An agent that reads a code base, edits files, runs commands, and opens pull requests. It runs in the terminal, VS Code, JetBrains, the desktop app, and the browser.
Agent tools: Claude Cowork · desktop, web, phone
Runs multi-step tasks in the folders, apps, and browser you allow, including scheduled tasks. Local file access needs the desktop app.
Claude in Chrome · browser
A Chrome extension that reads the open page and clicks, types, and fills forms across tabs. It is on paid plans.
Developer platform: Claude Platform · API
The developer API and Console for calling Claude models, with batch processing, prompt caching, tool use, and client SDKs.
Claude Managed Agents · API
A hosted agent harness that runs Claude with tools, MCP servers, and skills in a cloud or self-hosted sandbox. It is in beta.
For organizations: Claude Team and Enterprise · web, desktop, phone
Business plans with admin controls and SSO; Enterprise adds SCIM, audit logs, and a Compliance API.
Images, video, and audio: Claude Design · web, terminal
Builds mockups, one-pagers, and other visual assets from a description, with export to PPTX, PDF, or HTML. Claude Motion, in beta on Team and Enterprise, turns content into short animations.

Models

Claude Fable 5.1Current, Sep 2026
For demanding reasoning and long-horizon agent work.
1M context · reads text, image · writes text · $10.00 in, $50.00 out per million tokens
- Claude Fable 5.1 (company page)
- Pricing (company page)
Claude Opus 5.5Current, Sep 2026
Suggested starting model for most work, built for long-running agentic coding and knowledge work.
1M context · reads text, image · writes text · $4.00 in, $20.00 out per million tokens
- Claude Opus 5.5 (company page)
- Models overview (company page)
- Pricing (company page)
Claude Sonnet 5.5Current, Sep 2026
Combines speed and capability.
1M context · reads text, image · writes text · $2.00 in, $10.00 out per million tokens
- Claude Sonnet 5.5 (company page)
- Pricing (company page)
Claude Haiku 5.5Current, Oct 2026
Fastest model, for high-volume, latency-sensitive tasks such as classification and routing.
1M context · reads text, image · writes text · $0.10 in, $0.50 out per million tokens
- Claude Haiku 5.5 (company page)
- Pricing (company page)
Claude Fable 5Previous, Jun 2026
Earlier Fable model, still offered as a legacy option.
1M context · reads text, image · writes text · $10.00 in, $50.00 out per million tokens
- Claude Fable 5 (company page)
- Pricing (company page)
Claude Opus 5Previous, Jul 2026
Earlier Opus model, still offered as a legacy option.
1M context · reads text, image · writes text · $5.00 in, $25.00 out per million tokens
- Claude Opus 5 (company page)
- Pricing (company page)
Claude Sonnet 5Previous, Jun 2026
Earlier Sonnet model, still offered as a legacy option.
1M context · reads text, image · writes text · $2.00 in, $10.00 out per million tokens
- Claude Sonnet 5 (company page)
- Pricing (company page)

Capabilities

Apps

OfferedChat app: Claude
OfferedDeep research: ResearchPaid plans only.
LimitedLive voice: Voice modeBeta on all plans; languages other than English are also in beta.
Not offeredImage generationMakes diagrams and charts in HTML and SVG, not photos or illustrations.
Not offeredVideo generationClaude Motion exports code-built animations as MP4; it has no video model.
OfferedMemory: MemoryOn by default for personal plans; admins control it on Team and Enterprise.

Build

OfferedModel API: Claude API
Not offeredOpen weightsModels are offered only through the API and partner clouds.
OfferedBatch discount: Batch API50% off input and output.
OfferedPrompt caching: Prompt caching
Not offeredFine-tuningNot in the Claude API docs; past Claude fine-tuning ran on Amazon Bedrock, a partner.

Agents

OfferedCoding agent: Claude Code
OfferedAgent SDK: Claude Agent SDKPython and TypeScript.
OfferedComputer use: Computer use tool, Claude in Chrome
OfferedMCP support: Claude Code, Agent SDK, Managed Agents
OfferedHooks: Claude Code hooks
OfferedPlugins: Claude Code plugins
OfferedSkills: Agent Skills
OfferedSubagents: Claude Code subagents

Business

OfferedBusiness plan: Team, Enterprise
OfferedNo training by default: Commercial terms for API, Team, and EnterpriseUnless the customer joins the Development Partner Program or sends feedback.
OfferedMajor clouds: Amazon Bedrock, Google Cloud, Microsoft Foundry
LimitedData residency: inference_geoAPI only; US-only inference at 1.1x price, and US is the only storage region.

All 35 sources

Models overview (company page)
Claude API features overview (company page)
Pricing (company page)
Claude Opus 5.5 (company page)
Claude Sonnet 5.5 (company page)
Claude Platform (company page)
Anthropic privacy policy (registered address) (company page)
Claude Fable 5.1 (company page)
Claude Haiku 5.5 (company page)
Claude Fable 5 (company page)
Claude Opus 5 (company page)
Claude Sonnet 5 (company page)
Claude plans and pricing (company page)
Claude product overview (company page)
Claude Code overview (company page)
Claude Code hooks (company page)
Claude Code subagents (company page)
Claude Code skills (company page)
Claude Code MCP (company page)
Claude Code plugins (company page)
Agent SDK overview (company page)
Claude Managed Agents overview (company page)
Computer use tool (company page)
Skills (company page)
Claude Cowork (company page)
Claude in Chrome (company page)
Claude Design (company page)
Claude Motion help article (company page)
Using voice mode (company page)
Using the research feature (company page)
Claude takes research to new places (company page)
Using memory (company page)
Can Claude produce images? (company page)
How Anthropic uses data in model training (company page)
Data residency (company page)

How This Page Is Checked

Each fact comes from the company's own documentation, pricing page, or announcement, and each company records the date it was last checked. A mark of Limited means the feature is in preview, tied to a plan or region, or found in only one product; the cell's note says which.

Benchmark scores are left out on purpose. Most are the maker's own numbers, run on tests the maker chose. To read them well, see how to read a model claim. For launches as they happen, the AI news page labels each story by what its evidence shows.

The facts are refreshed about monthly and after major launches. Something out of date? The sources are listed in each profile, so you can check the current page yourself.