A/GEOResearch

Directory · 65 entries

Directory of AI search tools and AI labs

Find the tools that track and improve AI search visibility, and the AI companies behind AI answers: model labs, decision models, agent memory and reinforcement learning startups. Select an entry to see what it does, how you use it and its sources.

65 entries

ProfoundAI visibility trackers tryprofound.com Enterprise platform for answer engine insights, prompt tracking and AI agent analytics. AppAPIMCPPaid

Profound

Enterprise platform for answer engine insights, prompt tracking and AI agent analytics.

AI visibility trackersPaid

What is Profound?

Profound tracks how AI answer engines mention and cite a brand across a set of prompts. It also shows which AI agents and crawlers visit your site. Large brands and agencies use it for a full AEO and GEO program.

What you can do with it

  • Track citation share for your prompts against competitors
  • See which pages AI crawlers read
  • Connect the data to Claude or other AI tools through MCP

How you use it

AppAPIMCP

Quick reference

Category
AI visibility trackers
Pricing
Paid
Starting price (reported)
$99 per month, billed yearly
Best for
Dedicated AEO and GEO programs

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Peec AIAI visibility trackers peec.ai AI search analytics for marketing teams: visibility, position and sentiment. AppFree trial

Peec AI

AI search analytics for marketing teams: visibility, position and sentiment.

AI visibility trackersFree trial

What is Peec AI?

Peec AI tracks brand visibility, position and sentiment in AI answers for a set of prompts. It suggests prompts to track and compares you with competitors. Teams that work in many languages use it.

What you can do with it

  • Measure visibility, position and sentiment per prompt
  • Find the sources that AI answers cite in your category
  • Share reports with clients

How you use it

App

Quick reference

Category
AI visibility trackers
Pricing
Free trial
Starting price (reported)
€89 per month
Best for
Mid-market and multi-language teams

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

ScrunchAI visibility trackers scrunch.com AI visibility monitoring plus an agent experience layer for AI crawlers. AppPaid

Scrunch

AI visibility monitoring plus an agent experience layer for AI crawlers.

AI visibility trackersPaid

What is Scrunch?

Scrunch monitors how AI answer engines describe and cite a brand. Its Agent Experience Platform (AXP) serves AI crawlers a version of your pages that they can read more easily.

What you can do with it

  • Monitor brand answers across several AI engines
  • Find pages that AI crawlers misread
  • Serve AI agents a cleaner copy of your content

How you use it

App

Quick reference

Category
AI visibility trackers
Pricing
Paid
Starting price (reported)
$250 per month, billed yearly
Best for
Sites that AI crawlers misread

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Otterly.AIAI visibility trackers otterly.ai Low-cost AI search monitoring with GEO audits and prompt research. AppFree trial

Otterly.AI

Low-cost AI search monitoring with GEO audits and prompt research.

AI visibility trackersFree trial

What is Otterly.AI?

Otterly.AI tracks brand mentions and website citations in AI answers for the prompts you choose. It also runs GEO audits on your pages. Small teams use it as a first tracker.

What you can do with it

  • Track a small prompt set across four AI engines
  • Run a GEO audit on a page
  • Research the prompts that buyers ask

How you use it

App

Quick reference

Category
AI visibility trackers
Pricing
Free trial
Starting price (reported)
$29 per month
Engines included (reported)
4, with more as add-ons

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

AthenaHQAI visibility trackers athenahq.ai GEO platform that tracks AI citations and the sources AI tools trust. AppPaid

AthenaHQ

GEO platform that tracks AI citations and the sources AI tools trust.

AI visibility trackersPaid

What is AthenaHQ?

AthenaHQ tracks how AI answer engines cite a brand and which sources they trust. It suggests content actions to close the gaps.

What you can do with it

  • Track citations by prompt and engine
  • Map the sources that AI answers trust
  • Get a list of content actions

How you use it

App

Quick reference

Category
AI visibility trackers
Pricing
Paid
Best for
Citation and source analysis

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Semrush AI Visibility ToolkitAI visibility trackers semrush.com AI visibility data inside the Semrush SEO suite. AppAPIPaid

Semrush AI Visibility Toolkit

AI visibility data inside the Semrush SEO suite.

AI visibility trackersPaid

What is Semrush AI Visibility Toolkit?

The Semrush AI Visibility Toolkit adds AI answer tracking to the Semrush SEO platform. Teams that already use Semrush can see SEO and AI search data in one place.

What you can do with it

  • See AI visibility next to keyword rankings
  • Compare AI mentions with competitors
  • Find the prompts in your market

How you use it

AppAPI

Quick reference

Category
AI visibility trackers
Pricing
Paid
Best for
Teams that already use Semrush

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Ahrefs Brand RadarAI visibility trackers ahrefs.com Brand visibility in AI Overviews and AI chats, from the Ahrefs SEO platform. AppAPIMCPPaid

Ahrefs Brand Radar

Brand visibility in AI Overviews and AI chats, from the Ahrefs SEO platform.

AI visibility trackersPaid

What is Ahrefs Brand Radar?

Ahrefs Brand Radar tracks brand mentions in AI Overviews and AI chat tools. Ahrefs also has a hosted MCP server, so AI agents can query the data with your API units.

What you can do with it

  • Compare brand visibility in AI Overviews
  • Find content gaps against competitors
  • Query the data from an AI assistant through MCP

How you use it

AppAPIMCP

Quick reference

Category
AI visibility trackers
Pricing
Paid
Best for
Teams that already use Ahrefs

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

HubSpot AEOAI visibility trackers hubspot.com AI visibility tracking connected to HubSpot CRM data. AppFree trial

HubSpot AEO

AI visibility tracking connected to HubSpot CRM data.

AI visibility trackersFree trial

What is HubSpot AEO?

HubSpot AEO tracks how ChatGPT, Gemini and Perplexity answer prompts about your brand. It connects the data to your HubSpot CRM, so you can decide which gaps matter most.

What you can do with it

  • Track AI answers for your brand prompts
  • Link AI visibility to CRM data
  • Run the free AI Search Grader once

How you use it

App

Quick reference

Category
AI visibility trackers
Pricing
Free trial
Starting price (reported)
$50 per month
Engines (reported)
ChatGPT, Gemini, Perplexity

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

KnowatoaAI visibility trackers knowatoa.com AI search tracking for agencies that manage many brands. AppFree trial

Knowatoa

AI search tracking for agencies that manage many brands.

AI visibility trackersFree trial

What is Knowatoa?

Knowatoa tracks how AI tools answer questions about a brand and its competitors. Agencies use it to report on many brands from one account.

What you can do with it

  • Track AI answers for several client brands
  • Compare brands in one report
  • Watch changes over time

How you use it

App

Quick reference

Category
AI visibility trackers
Pricing
Free trial
Starting price (reported)
$59 per month
Best for
Agencies with many brands

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

PromptwatchAI visibility trackers promptwatch.com AI search tracking across many AI platforms, built for agencies. AppPaid

Promptwatch

AI search tracking across many AI platforms, built for agencies.

AI visibility trackersPaid

What is Promptwatch?

Promptwatch tracks brand mentions and citations across many AI platforms. Agencies use it to monitor many brands at the same time.

What you can do with it

  • Track mentions on many AI platforms
  • Monitor several client brands
  • Export reports for clients

How you use it

App

Quick reference

Category
AI visibility trackers
Pricing
Paid
Best for
Agencies

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Writesonic GEOAI visibility trackers writesonic.com AI visibility tracking with built-in content writing. AppPaid

Writesonic GEO

AI visibility tracking with built-in content writing.

AI visibility trackersPaid

What is Writesonic GEO?

Writesonic GEO tracks how AI answers cite your brand and helps you write content to close the gaps. It joins tracking and content work in one tool.

What you can do with it

  • Track AI citations for your prompts
  • Draft content for prompts you miss
  • Check competitor mentions

How you use it

App

Quick reference

Category
AI visibility trackers
Pricing
Paid
Best for
Teams that want tracking and writing in one tool

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

A/GEO AI Visibility ScoreAI visibility trackers ageoresearch.com/ai-visibility-score · by A/GEO Research Free scan: can AI tools read and cite your site? 37 checks in about 20 seconds. AppFree

A/GEO AI Visibility Score

Free scan: can AI tools read and cite your site? 37 checks in about 20 seconds.

AI visibility trackersFree

What is A/GEO AI Visibility Score?

The AI Visibility Score checks if AI tools can read and cite a website. It runs 37 checks on the home page, robots.txt, the sitemap and llms.txt. It measures readiness, not how often AI tools cite you.

What you can do with it

  • Get a score out of 100 for any site
  • See the three fixes that add the most points
  • Check a client site before a pitch

How you use it

App

Quick reference

Category
AI visibility trackers
Pricing
Free
Checks
37
Time
About 20 seconds
Made by
A/GEO Research

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Google AI Mode and AI OverviewsAI answer engines search.google.com AI answers inside Google Search, built on the Google Search index. AppFree

Google AI Mode and AI Overviews

AI answers inside Google Search, built on the Google Search index.

AI answer enginesFree

What is Google AI Mode and AI Overviews?

AI Overviews show an AI summary above the search results. AI Mode gives a full AI answer with links. Both use pages from the Google Search index, so normal SEO rules apply. Google-Extended is a control token for training, not a crawler.

What you can do with it

  • Check which pages Google cites for your queries
  • Keep pages indexable by Googlebot
  • Follow Google’s guide for AI features

How you use it

App

Quick reference

Category
AI answer engines
Pricing
Free
Company
Google
Crawler
Googlebot (Search index)

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

GeminiAI answer engines gemini.google.com Google’s AI assistant app, with answers grounded in Google Search. AppAPIFree tier

Gemini

Google’s AI assistant app, with answers grounded in Google Search.

AI answer enginesFree tier

What is Gemini?

Gemini is Google’s AI assistant. It can ground answers in Google Search results and cite web pages. The same Gemini models power AI Mode.

What you can do with it

  • Ask your buyer questions in the Gemini app
  • Compare Gemini answers with AI Mode
  • Track the sources it cites

How you use it

AppAPI

Quick reference

Category
AI answer engines
Pricing
Free tier
Company
Google

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

PerplexityAI answer engines perplexity.ai AI answer engine that searches the web for every answer and shows its sources. AppAPIFree tier

Perplexity

AI answer engine that searches the web for every answer and shows its sources.

AI answer enginesFree tier

What is Perplexity?

Perplexity searches the web for each question and writes an answer with numbered sources. PerplexityBot finds pages for the index, and Perplexity-User fetches pages when a person asks.

What you can do with it

  • Ask your buyer questions and check the sources list
  • Allow PerplexityBot in robots.txt
  • Track citation changes from week to week

How you use it

AppAPI

Quick reference

Category
AI answer engines
Pricing
Free tier
Crawlers
PerplexityBot, Perplexity-User

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Microsoft CopilotAI answer engines copilot.microsoft.com Microsoft’s AI assistant, with web answers grounded in Bing. AppFree tier

Microsoft Copilot

Microsoft’s AI assistant, with web answers grounded in Bing.

AI answer enginesFree tier

What is Microsoft Copilot?

Microsoft Copilot answers questions with web results from Bing and cites its sources. Pages that Bing can index are the pages Copilot can cite.

What you can do with it

  • Submit your sitemap in Bing Webmaster Tools
  • Use IndexNow so Bing sees changes fast
  • Ask your buyer questions in Copilot

How you use it

App

Quick reference

Category
AI answer engines
Pricing
Free tier
Company
Microsoft
Crawler
Bingbot (Bing index)

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

ClaudeAI answer engines claude.ai Anthropic’s AI assistant. It can search the web and cite pages. AppAPIMCPFree tier

Claude

Anthropic’s AI assistant. It can search the web and cite pages.

AI answer enginesFree tier

What is Claude?

Claude is Anthropic’s AI assistant. With web search on, it finds and cites web pages. Claude-SearchBot finds pages for search, Claude-User fetches a page when a person asks, and ClaudeBot collects data for training.

What you can do with it

  • Ask your buyer questions with web search on
  • Allow Claude-SearchBot to be found in search
  • Decide on ClaudeBot (training) as a separate choice

How you use it

AppAPIMCP

Quick reference

Category
AI answer engines
Pricing
Free tier
Company
Anthropic
Crawlers
ClaudeBot, Claude-SearchBot, Claude-User

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

GrokAI answer engines grok.com xAI’s AI assistant, with answers from the web and from X. AppAPIFree tier

Grok

xAI’s AI assistant, with answers from the web and from X.

AI answer enginesFree tier

What is Grok?

Grok is the AI assistant from xAI. It answers with results from the web and from posts on X.

What you can do with it

  • Ask your buyer questions in Grok
  • Check how posts on X change the answer
  • Track the sources it cites

How you use it

AppAPI

Quick reference

Category
AI answer engines
Pricing
Free tier
Company
xAI
Crawler
GrokBot

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Meta AIAI answer engines meta.ai Meta’s AI assistant in WhatsApp, Instagram, Facebook and the web. AppFree

Meta AI

Meta’s AI assistant in WhatsApp, Instagram, Facebook and the web.

AI answer enginesFree

What is Meta AI?

Meta AI is the AI assistant in Meta apps. People ask it for recommendations inside WhatsApp, Instagram and Facebook. Meta-ExternalAgent is Meta’s crawler.

What you can do with it

  • Ask your buyer questions in Meta AI
  • Decide on Meta-ExternalAgent in robots.txt
  • Keep your Facebook and Instagram facts current

How you use it

App

Quick reference

Category
AI answer engines
Pricing
Free
Company
Meta
Crawler
Meta-ExternalAgent

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Cloudflare AI Crawl ControlAI crawler control developers.cloudflare.com See which AI crawlers visit your site, set a rule for each one, or charge them per crawl. AppAPIFree tier

Cloudflare AI Crawl Control

See which AI crawlers visit your site, set a rule for each one, or charge them per crawl.

AI crawler controlFree tier

What is Cloudflare AI Crawl Control?

AI Crawl Control (formerly AI Audit) shows which AI services read your content on Cloudflare. You can allow or block each crawler, track robots.txt compliance and test pay per crawl. Free plans show the last 24 hours of data.

What you can do with it

  • See AI crawler visits by crawler, path and status code
  • Block or allow each AI crawler
  • Query the data with the GraphQL Analytics API

How you use it

AppAPI

Quick reference

Category
AI crawler control
Pricing
Free tier
Needs
A domain that proxies through Cloudflare
Free plan data
Last 24 hours

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Known Agents (formerly Dark Visitors)AI crawler control knownagents.com Keeps robots.txt current as new AI agents launch, and logs AI crawler visits. AppAPIFree tier

Known Agents (formerly Dark Visitors)

Keeps robots.txt current as new AI agents launch, and logs AI crawler visits.

AI crawler controlFree tier

What is Known Agents?

Known Agents, formerly Dark Visitors, keeps a list of AI agents and crawlers. It can update your robots.txt as new agents launch and show you which agents visit your site.

What you can do with it

  • Keep robots.txt rules current for new AI agents
  • See AI crawler visits in real time
  • Track referrals from AI tools

How you use it

AppAPI

Quick reference

Category
AI crawler control
Pricing
Free tier
Former name
Dark Visitors

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

TollBitAI crawler control tollbit.com Monitor, manage and charge AI bots that read your content. AppContact sales

TollBit

Monitor, manage and charge AI bots that read your content.

AI crawler controlContact sales

What is TollBit?

TollBit helps publishers see AI bot traffic and license their content to AI companies. Its model is monitor, manage and monetize.

What you can do with it

  • Measure AI bot scrapes of your content
  • Set terms for AI access
  • License content to AI companies

How you use it

App

Quick reference

Category
AI crawler control
Pricing
Contact sales
Best for
Publishers

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

AirOpsContent and schema airops.com Content engineering platform for AI search visibility and content operations. AppAPIMCPFree tier

AirOps

Content engineering platform for AI search visibility and content operations.

Content and schemaFree tier

What is AirOps?

AirOps helps teams plan, write and refresh content at scale with AI workflows. It focuses on content that AI search tools cite.

What you can do with it

  • Build content workflows with AI steps
  • Refresh old pages for AI search
  • Track content performance

How you use it

AppAPIMCP

Quick reference

Category
Content and schema
Pricing
Free tier
Best for
Content teams at scale

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Schema AppContent and schema schemaapp.com Enterprise schema markup that builds a content knowledge graph. AppContact sales

Schema App

Enterprise schema markup that builds a content knowledge graph.

Content and schemaContact sales

What is Schema App?

Schema App writes and manages schema.org markup across large websites. It connects your pages into a knowledge graph that search engines and AI tools can read.

What you can do with it

  • Add schema markup at scale
  • Build a knowledge graph of your entities
  • Keep markup current when pages change

How you use it

App

Quick reference

Category
Content and schema
Pricing
Contact sales
Best for
Large websites

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

WordLiftContent and schema wordlift.io Knowledge graph and structured data for WordPress and other sites. PluginAPIPaid

WordLift

Knowledge graph and structured data for WordPress and other sites.

Content and schemaPaid

What is WordLift?

WordLift builds a knowledge graph from your content and adds structured data to your pages. It has a WordPress plugin and works with other platforms too.

What you can do with it

  • Mark up entities on your pages
  • Build a knowledge graph of your site
  • Link your content to public knowledge bases

How you use it

PluginAPI

Quick reference

Category
Content and schema
Pricing
Paid
Platforms
WordPress and others

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Yoast SEOContent and schema yoast.com WordPress SEO plugin with schema output and an llms.txt option. PluginFree tier

Yoast SEO

WordPress SEO plugin with schema output and an llms.txt option.

Content and schemaFree tier

What is Yoast SEO?

Yoast SEO adds meta tags, sitemaps and schema markup to WordPress sites. It can also make an llms.txt file for AI tools.

What you can do with it

  • Add schema and meta tags to WordPress pages
  • Make an XML sitemap
  • Make an llms.txt file

How you use it

Plugin

Quick reference

Category
Content and schema
Pricing
Free tier
Platform
WordPress

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

A/GEO Schema ManagerContent and schema ageoresearch.com/ageo-plugin-wp · by A/GEO Research Free WordPress plugin to add JSON-LD schema and meta tags per page. PluginFree

A/GEO Schema Manager

Free WordPress plugin to add JSON-LD schema and meta tags per page.

Content and schemaFree

What is A/GEO Schema Manager?

A/GEO Schema Manager adds JSON-LD schema and meta tags at the page, post and custom post type level. It has a dashboard to manage the code for the whole site.

What you can do with it

  • Paste JSON-LD for one page or the whole site
  • Use code from the free Schema Generator
  • Manage meta tags without a theme change

How you use it

Plugin

Quick reference

Category
Content and schema
Pricing
Free
Platform
WordPress
Made by
A/GEO Research

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

OpenAIModel labs developers.openai.com Flagship: GPT-6 Astra. App: ChatGPT. Weights: closed. APIAppPaid API

OpenAI

Flagship: GPT-6 Astra. App: ChatGPT. Weights: closed.

Model labsPaid API

What is OpenAI?

Four GPT-6 models share a 1.05M-token context window. Astra is the most capable. GPT-6.1 Sol gives near-Astra results at one fifth of the price. GPT-6.1 Terra costs half as much as Sol and is made for computer use. Luna is the low-cost tier. Two special models are for approved customers only: GPT-6 Rosalind (life sciences, October 30) and GPT-6 Cyber (security, November 5). On November 10, OpenAI opened Businesses in ChatGPT (beta): a business can claim a verified profile, change its details and see how often ChatGPT mentions it.

What you can do with it

  • Compare OpenAI models, prices and context windows on LLM Watch
  • Set robots.txt rules for GPTBot, OAI-SearchBot, ChatGPT-User
  • Call its models through the API

How you use it

APIApp

Quick reference

Category
Model labs
Pricing
Paid API
Flagship model
GPT-6 Astra
Country
US
Weights
Closed
Intelligence Index
57 (GPT-6 Astra (max))
Crawlers
GPTBot, OAI-SearchBot, ChatGPT-User
Newest model in LLM Watch
GPT-6 Cyber (2026-11-05)

Sources

From LLM Watch, checked November 13, 2026. Prices and funding change. Check the vendor site before you buy.

AnthropicModel labs platform.claude.com Flagship: Claude Fable 5.2. App: Claude. Weights: closed. APIAppPaid API

Anthropic

Flagship: Claude Fable 5.2. App: Claude. Weights: closed.

Model labsPaid API

What is Anthropic?

Four current public Claude models, all with a 1M-token context window. Claude Fable 5.2 (November 9) has the top score on the Artificial Analysis index (65). It costs less than Fable 5.1 and Anthropic says it writes output up to 30% faster. Anthropic still recommends Opus 5.5 for most work. Claude Mythos 5.2 is the same model with fewer safeguards, for verified organizations only.

What you can do with it

  • Compare Anthropic models, prices and context windows on LLM Watch
  • Set robots.txt rules for ClaudeBot, Claude-SearchBot, Claude-User
  • Call its models through the API

How you use it

APIApp

Quick reference

Category
Model labs
Pricing
Paid API
Flagship model
Claude Fable 5.2
Country
US
Weights
Closed
Intelligence Index
65 (Claude Fable 5.2 (max))
Crawlers
ClaudeBot, Claude-SearchBot, Claude-User
Newest model in LLM Watch
Claude Fable 5.2 (2026-11-09)

Sources

From LLM Watch, checked November 13, 2026. Prices and funding change. Check the vendor site before you buy.

GoogleModel labs deepmind.google Flagship: Gemini 4 Carbon. App: Gemini. Weights: closed. APIAppPaid API

Google

Flagship: Gemini 4 Carbon. App: Gemini. Weights: closed.

Model labsPaid API

What is Google?

Gemini 4 now has three models. Gemini 4 Carbon has the top Google score on the Artificial Analysis index (60). It opened to paying developers and AI Ultra users on November 10. Gemini 4 Hydrogen (November 12) is the fast, low-cost model, open to everyone. It is now the default model in AI Mode and answers the hardest questions in AI Overviews. The low intro prices for Carbon and Argon end on November 30.

What you can do with it

  • Compare Google models, prices and context windows on LLM Watch
  • Set robots.txt rules for Googlebot
  • Call its models through the API

How you use it

APIApp

Quick reference

Category
Model labs
Pricing
Paid API
Flagship model
Gemini 4 Carbon
Country
US
Weights
Closed
Intelligence Index
60 (Gemini 4 Carbon (high))
Crawlers
Googlebot (AI Overviews use the Search index); Google-Extended is a control token
Newest model in LLM Watch
Gemini 4 Hydrogen (2026-11-12)

Sources

From LLM Watch, checked November 13, 2026. Prices and funding change. Check the vendor site before you buy.

xAI (SpaceXAI)Model labs docs.x.ai Flagship: Grok 4.7. App: Grok. Weights: closed. APIAppPaid API

xAI (SpaceXAI)

Flagship: Grok 4.7. App: Grok. Weights: closed.

Model labsPaid API

What is xAI?

Grok 4.7 is the current API model, with a 500K-token context window. Cursor co-develops Grok models with SpaceXAI from Grok 4.5 onward.

What you can do with it

  • Compare xAI models, prices and context windows on LLM Watch
  • Set robots.txt rules for GrokBot
  • Call its models through the API

How you use it

APIApp

Quick reference

Category
Model labs
Pricing
Paid API
Flagship model
Grok 4.7
Country
US
Weights
Closed
Intelligence Index
35 (Grok 4.7 (xhigh))
Crawlers
GrokBot
Newest model in LLM Watch
Grok 4.7 (2026-09-21)

Sources

From LLM Watch, checked November 13, 2026. Prices and funding change. Check the vendor site before you buy.

MetaModel labs dev.meta.ai Flagship: Muse Wonder Full. App: Meta AI. Weights: mixed. APIAppOpen weightsPaid API

Meta

Flagship: Muse Wonder Full. App: Meta AI. Weights: mixed.

Model labsPaid API

What is Meta?

Meta moved from open Llama models to the Muse line. Muse Wonder (October 12) is on the Meta Model API. Muse Wonder Full is only in Muse Code. Muse Ember (November 9) is the smallest and lowest-cost Muse model, for high-volume agent tasks. Muse Spark Open (October 15) and Muse Glimmer 30B have open weights.

What you can do with it

  • Compare Meta models, prices and context windows on LLM Watch
  • Set robots.txt rules for Meta-ExternalAgent
  • Download open weights and run a model yourself

How you use it

APIAppOpen weights

Quick reference

Category
Model labs
Pricing
Paid API
Flagship model
Muse Wonder Full
Country
US
Weights
Mixed
Intelligence Index
50 (Muse Wonder Full (max))
Crawlers
Meta-ExternalAgent
Newest model in LLM Watch
Muse Ember (2026-11-09)

Sources

From LLM Watch, checked November 13, 2026. Prices and funding change. Check the vendor site before you buy.

DeepSeekModel labs api-docs.deepseek.com Flagship: DeepSeek V4-Pro. App: DeepSeek. Weights: open. APIAppOpen weightsOpen weights

DeepSeek

Flagship: DeepSeek V4-Pro. App: DeepSeek. Weights: open.

Model labsOpen weights

What is DeepSeek?

Two API models, both with a 1M-token context window. Prices drop by half outside peak hours (peak: 01:00 to 04:00 and 06:00 to 10:00 UTC, weekdays).

What you can do with it

  • Compare DeepSeek models, prices and context windows on LLM Watch
  • Test its models through the API
  • Download open weights and run a model yourself

How you use it

APIAppOpen weights

Quick reference

Category
Model labs
Pricing
Open weights
Flagship model
DeepSeek V4-Pro
Country
China
Weights
Open
Intelligence Index
33 (DeepSeek V4.1 Flash 1017 (max))
Newest model in LLM Watch
DeepSeek V4.1 Flash 1017 (2026-10-17)

Sources

From LLM Watch, checked November 13, 2026. Prices and funding change. Check the vendor site before you buy.

Alibaba (Qwen)Model labs qwen.ai Flagship: Qwen3.8-Max. App: Qwen Chat. Weights: mixed. APIAppOpen weightsPaid API

Alibaba (Qwen)

Flagship: Qwen3.8-Max. App: Qwen Chat. Weights: mixed.

Model labsPaid API

What is Alibaba?

Qwen3.8-Max has 2.4 trillion parameters, with about 95 billion active for each token. The first Qwen 4 model, Qwen4-Max-Preview, opened on November 11 in the Qwen app and on Alibaba's Token Plan subscription only. The preview ends on November 25. Alibaba has not published a price, benchmarks or weights for Qwen 4.

What you can do with it

  • Compare Alibaba models, prices and context windows on LLM Watch
  • Test its models through the API
  • Download open weights and run a model yourself

How you use it

APIAppOpen weights

Quick reference

Category
Model labs
Pricing
Paid API
Flagship model
Qwen3.8-Max
Country
China
Weights
Mixed
Intelligence Index
34 (Qwen3.8 Max (0902))
Newest model in LLM Watch
Qwen4-Max-Preview (2026-11-11)

Sources

From LLM Watch, checked November 13, 2026. Prices and funding change. Check the vendor site before you buy.

Moonshot AI (Kimi)Model labs platform.moonshot.ai Flagship: Kimi K3.1 Pro. App: Kimi. Weights: open. APIAppOpen weightsOpen weights

Moonshot AI (Kimi)

Flagship: Kimi K3.1 Pro. App: Kimi. Weights: open.

Model labsOpen weights

What is Moonshot AI?

Kimi K3.1 Pro (October 23) is the new flagship, with a 1M-token context window. It is second on EQ-Bench Creative Writing v3. Moonshot posted its open weights on Hugging Face on November 4, 2026, as it promised. The files are about 1.56 TB. Kimi K3 (2.8 trillion parameters) has had open weights since July 26, 2026.

What you can do with it

  • Compare Moonshot AI models, prices and context windows on LLM Watch
  • Test its models through the API
  • Download open weights and run a model yourself

How you use it

APIAppOpen weights

Quick reference

Category
Model labs
Pricing
Open weights
Flagship model
Kimi K3.1 Pro
Country
China
Weights
Open
Intelligence Index
46 (Kimi K3.1 Pro)
Newest model in LLM Watch
Kimi K3.1 Pro (2026-10-23)

Sources

From LLM Watch, checked November 13, 2026. Prices and funding change. Check the vendor site before you buy.

Z.ai (GLM)Model labs z.ai Flagship: GLM-5.3. App: Z.ai chat. Weights: open. APIAppOpen weightsOpen weights

Z.ai (GLM)

Flagship: GLM-5.3. App: Z.ai chat. Weights: open.

Model labsOpen weights

What is Z.ai?

GLM-5.3 uses the GLM-5.2 base model with new post-training. GLM-5.3-Flash is the low-cost open model under the MIT license.

What you can do with it

  • Compare Z.ai models, prices and context windows on LLM Watch
  • Test its models through the API
  • Download open weights and run a model yourself

How you use it

APIAppOpen weights

Quick reference

Link
z.ai
Category
Model labs
Pricing
Open weights
Flagship model
GLM-5.3
Country
China
Weights
Open
Intelligence Index
32 (GLM-5.3 (max))
Newest model in LLM Watch
GLM-5.3-Flash (2026-08-26)

Sources

From LLM Watch, checked November 13, 2026. Prices and funding change. Check the vendor site before you buy.

Mistral AIModel labs mistral.ai Flagship: Mistral Large 4. App: Le Chat. Weights: open. APIAppOpen weightsOpen weights

Mistral AI

Flagship: Mistral Large 4. App: Le Chat. Weights: open.

Model labsOpen weights

What is Mistral AI?

Mistral Large 4 has about 1.05 trillion parameters, with 52 billion active. Mistral posted the open weights on October 29, 2026, under the Mistral Ecosystem Open-Weight License (based on MIT). A company that sells Large 4 as a hosted service needs a separate license from Mistral.

What you can do with it

  • Compare Mistral AI models, prices and context windows on LLM Watch
  • Set robots.txt rules for MistralAI-User
  • Download open weights and run a model yourself

How you use it

APIAppOpen weights

Quick reference

Category
Model labs
Pricing
Open weights
Flagship model
Mistral Large 4
Country
France
Weights
Open
Intelligence Index
26 (Mistral Large 4 (high))
Crawlers
MistralAI-User
Newest model in LLM Watch
Mistral Large 4 (2026-10-06)

Sources

From LLM Watch, checked November 13, 2026. Prices and funding change. Check the vendor site before you buy.

Xiaomi (MiMo)Model labs mimo.xiaomi.com Flagship: MiMo-V2.6-Pro. Weights: open. APIOpen weightsOpen weights

Xiaomi (MiMo)

Flagship: MiMo-V2.6-Pro. Weights: open.

Model labsOpen weights

What is Xiaomi?

MiMo-V2.6-Pro is one of the strongest open-weight models on the Artificial Analysis index, at about $0.13 for each index task. Xiaomi retired MiMo-V2.5 on October 21.

What you can do with it

  • Compare Xiaomi models, prices and context windows on LLM Watch
  • Test its models through the API
  • Download open weights and run a model yourself

How you use it

APIOpen weights

Quick reference

Category
Model labs
Pricing
Open weights
Flagship model
MiMo-V2.6-Pro
Country
China
Weights
Open
Intelligence Index
30 (MiMo-V2.6-Pro)
Newest model in LLM Watch
MiMo-V2.6-Pro (2026-09-21)

Sources

From LLM Watch, checked November 13, 2026. Prices and funding change. Check the vendor site before you buy.

TypeSafe AI (Jev)Decision models rits.shanghai.nyu.edu Jev, a closed decision model for agents. Released Jev 1.14 on October 14, 2026. APIPaid

TypeSafe AI (Jev)

Jev, a closed decision model for agents. Released Jev 1.14 on October 14, 2026.

Decision modelsPaid

What is TypeSafe AI?

TypeSafe AI makes Jev, a decision model. Jev reads an input and returns one typed answer with a probability. Agents use it to route requests and flag problems. Our site chat uses Jev as an optional gate.

What you can do with it

  • Route a request to the right agent or reply
  • Flag off-topic or unsafe input before a large model runs
  • Score intent and buying stage

How you use it

API

Quick reference

Category
Decision models
Pricing
Paid
Latest model
Jev 1.14 (Oct 14, 2026)
Input price
$0.042 per 1M tokens, output free
Published score
JevBench 72.13 for Jev 1.13 (3rd of 112)

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Liquid AI (d1)Decision models liquid.ai d1-ultra and the open d1-3B decision models. APIOpen weightsPaid

Liquid AI (d1)

d1-ultra and the open d1-3B decision models.

Decision modelsPaid

What is Liquid AI?

Liquid AI makes the d1 decision models. d1-ultra is a hosted model. d1-3B has open weights under the LFM Open License, so you can run it yourself.

What you can do with it

  • Run fast yes or no checks in an agent
  • Self-host d1-3B for private data
  • Compare with Jev on your own tests

How you use it

APIOpen weights

Quick reference

Category
Decision models
Pricing
Paid
Models
d1-ultra (Oct 15, 2026), d1-3B (Oct 7, 2026)
d1-ultra input price
$0.042 per 1M tokens
d1-3B score
Decision Index 48.57 (self-reported)

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Base Labs (mammoth-1)Decision models aicoder.com Open decision models mammoth-1 (27B) and elephant-1 under the MIT license. Open weightsOpen source

Base Labs (mammoth-1)

Open decision models mammoth-1 (27B) and elephant-1 under the MIT license.

Decision modelsOpen source

What is Base Labs?

Base Labs released two open decision models, mammoth-1 and elephant-1, on October 21 and 22, 2026. The weights use the MIT license.

What you can do with it

  • Self-host a decision model with an MIT license
  • Fine-tune it on your own labels
  • Compare open and closed decision models

How you use it

Open weights

Quick reference

Category
Decision models
Pricing
Open source
mammoth-1 score
Jev Decision Index 71.24 (self-reported)
License
MIT

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Microsoft-Decision-1Decision models testingcatalog.com Microsoft’s decision model in Foundry, built on Qwen3.5-9B. APIPaid

Microsoft-Decision-1

Microsoft’s decision model in Foundry, built on Qwen3.5-9B.

Decision modelsPaid

What is Microsoft-Decision-1?

Microsoft-Decision-1 is a closed decision model in Microsoft Foundry. Microsoft released it on October 9, 2026 and has not published comparable scores.

What you can do with it

  • Add decision steps to agents built in Foundry
  • Classify and route requests at low cost
  • Test it against open models on your data

How you use it

API

Quick reference

Category
Decision models
Pricing
Paid
Released
Oct 9, 2026
Input price
$0.042 per 1M tokens, output free
Base
Qwen3.5-9B

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

OpenAI Decisions APIDecision models openai.com OpenAI’s decision endpoint on GPT-6 Luna, generally available since October 22, 2026. APIPaid

OpenAI Decisions API

OpenAI’s decision endpoint on GPT-6 Luna, generally available since October 22, 2026.

Decision modelsPaid

What is OpenAI Decisions API?

The Decisions API returns typed decisions from GPT-6 Luna. OpenAI says it answers up to 10 times faster than a chat call. OpenAI has not published comparable scores.

What you can do with it

  • Get typed yes or no and choice answers from an OpenAI model
  • Replace slow chat calls in agent routing
  • Keep one vendor for chat and decisions

How you use it

API

Quick reference

Category
Decision models
Pricing
Paid
Generally available
Oct 22, 2026
Input price
$0.10 per 1M tokens

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Cloudflare ClefDecision models infoq.com Open decision models Clef (27B) and Clef-flash (9B), Apache 2.0. APIOpen weightsPaid

Cloudflare Clef

Open decision models Clef (27B) and Clef-flash (9B), Apache 2.0.

Decision modelsPaid

What is Cloudflare Clef?

Cloudflare released the Clef decision models on October 1, 2026, with open weights under Apache 2.0. You can call them on Workers AI or run them yourself.

What you can do with it

  • Run decisions at the edge on Workers AI
  • Self-host with an Apache 2.0 license
  • Pick Clef-flash for the lowest cost

How you use it

APIOpen weights

Quick reference

Category
Decision models
Pricing
Paid
Clef input price
$0.24 per 1M tokens
Clef-flash input price
$0.038 per 1M tokens
Scores
Decision Index 61.2 and 57.1 (self-reported)

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Perplexity DeciderDecision models openrouter.ai Perplexity’s open 27B decision model, Apache 2.0. APIOpen weightsPaid

Perplexity Decider

Perplexity’s open 27B decision model, Apache 2.0.

Decision modelsPaid

What is Perplexity Decider?

Perplexity Decider v1.2 is a 27B decision model with open weights under Apache 2.0. Perplexity reports 85.71% on its own 11-task panel.

What you can do with it

  • Call it on OpenRouter at low cost
  • Self-host the open weights
  • Use it for search query routing

How you use it

APIOpen weights

Quick reference

Category
Decision models
Pricing
Paid
Input price
$0.02 per 1M tokens
Score
85.71% on its own 11-task panel

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Kakao Kanana-2-v-decideDecision models huggingface.co Open 3.67B decision model from Kakao, released November 6, 2026. Open weightsSelf-host

Kakao Kanana-2-v-decide

Open 3.67B decision model from Kakao, released November 6, 2026.

Decision modelsSelf-host

What is Kakao Kanana-2-v-decide?

Kakao released Kanana-2-v-decide, a 3.67B open decision model, on November 6, 2026. The weights use the Kanana Open License.

What you can do with it

  • Self-host a small decision model
  • Test decisions in Korean and English
  • Compare small open models

How you use it

Open weights

Quick reference

Category
Decision models
Pricing
Self-host
Released
Nov 6, 2026
Score
Decision Index average 61.23 (self-reported)

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Mem0Agent memory mem0.ai Memory layer for AI agents: open source library plus a managed platform. APISDKOpen sourceFree tier

Mem0

Memory layer for AI agents: open source library plus a managed platform.

Agent memoryFree tier

What is Mem0?

Mem0 extracts facts and preferences from conversations and finds them again for later sessions. You add it to an existing agent. It is open core under Apache 2.0, with a managed cloud API.

What you can do with it

  • Remember user preferences across sessions
  • Add memory to an agent you already run
  • Self-host the open source library

How you use it

APISDKOpen source

Quick reference

Category
Agent memory
Pricing
Free tier
Funding
$24M seed and Series A (Oct 2025)
BEAM 10M score
48.6% (self-reported)

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Letta (formerly MemGPT)Agent memory letta.com Stateful agent runtime where the agent edits its own memory. APISDKOpen sourceFree tier

Letta (formerly MemGPT)

Stateful agent runtime where the agent edits its own memory.

Agent memoryFree tier

What is Letta?

Letta runs agents that manage their own memory with tool calls. It came from the MemGPT research at UC Berkeley. It is open source under Apache 2.0, with Letta Cloud as the hosted option.

What you can do with it

  • Run long-running agents that keep state
  • Let the agent decide what to remember
  • Self-host the open source server

How you use it

APISDKOpen source

Quick reference

Category
Agent memory
Pricing
Free tier
Funding
$10M seed led by Felicis (Sep 2024)
Former name
MemGPT

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Zep (Graphiti)Agent memory getzep.com Temporal knowledge graph memory for agents. Graphiti is the open source core. APISDKOpen sourceFree tier

Zep (Graphiti)

Temporal knowledge graph memory for agents. Graphiti is the open source core.

Agent memoryFree tier

What is Zep?

Zep builds a knowledge graph from conversations and keeps track of when each fact was true. Graphiti is its open source graph engine. Zep Cloud is the hosted service.

What you can do with it

  • Answer questions about how facts changed over time
  • Give CRM and support agents a fact graph
  • Self-host Graphiti

How you use it

APISDKOpen source

Quick reference

Category
Agent memory
Pricing
Free tier
Approach
Temporal knowledge graph

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

SupermemoryAgent memory supermemory.ai Memory engine and API with many connectors, MIT licensed. APIOpen sourceFree tier

Supermemory

Memory engine and API with many connectors, MIT licensed.

Agent memoryFree tier

What is Supermemory?

Supermemory is a memory layer for AI apps. It reads many file types and data sources through connectors. The core engine and app are open source under the MIT license.

What you can do with it

  • Index documents and data sources into memory
  • Give an AI app a memory API
  • Self-host the MIT-licensed engine

How you use it

APIOpen source

Quick reference

Category
Agent memory
Pricing
Free tier
License
MIT (core engine)

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

CogneeAgent memory cognee.ai Agent memory as a knowledge graph built from your documents and data. SDKOpen sourceOpen source

Cognee

Agent memory as a knowledge graph built from your documents and data.

Agent memoryOpen source

What is Cognee?

Cognee builds agent memory as a knowledge graph from documents and other data, not only from chat history. It is open source.

What you can do with it

  • Turn documents into a knowledge graph
  • Give agents graph-based recall
  • Self-host the open source library

How you use it

SDKOpen source

Quick reference

Category
Agent memory
Pricing
Open source
Funding
$7.5M seed (2025)

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

past.devAgent memory past.dev Memory API with the top self-reported BEAM score in LLM Watch. APIPaid

past.dev

Memory API with the top self-reported BEAM score in LLM Watch.

Agent memoryPaid

What is past.dev?

past.dev sells a memory API for AI agents. It reports the highest BEAM scores in our LLM Watch table. No one has reproduced the scores yet.

What you can do with it

  • Add long-term memory through an API
  • Test recall on very long conversations
  • Compare with other memory layers on BEAM

How you use it

API

Quick reference

Category
Agent memory
Pricing
Paid
BEAM 10M score
85.03% (self-reported, Sep 29, 2026)

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Hindsight (Vectorize)Agent memory hindsight.vectorize.io Agent memory from Vectorize, with its own Agent Memory Benchmark. APIPaid

Hindsight (Vectorize)

Agent memory from Vectorize, with its own Agent Memory Benchmark.

Agent memoryPaid

What is Hindsight?

Hindsight is an agent memory system from Vectorize. Vectorize also runs the Agent Memory Benchmark site, where it reports BEAM scores.

What you can do with it

  • Add memory to agents
  • Compare memory systems on one benchmark site
  • Test recall at 1M and 10M tokens

How you use it

API

Quick reference

Category
Agent memory
Pricing
Paid
BEAM 10M score
64.1% (self-reported)

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Honcho (Plastic Labs)Agent memory plasticlabs.ai Memory and user modeling for AI apps from Plastic Labs. APIOpen sourceFree tier

Honcho (Plastic Labs)

Memory and user modeling for AI apps from Plastic Labs.

Agent memoryFree tier

What is Honcho?

Honcho from Plastic Labs stores what an AI app learns about each user and builds a model of that user over time.

What you can do with it

  • Build a profile of each user over time
  • Personalize answers in an AI app
  • Compare recall on BEAM

How you use it

APIOpen source

Quick reference

Category
Agent memory
Pricing
Free tier
BEAM 10M score
40.6% (Plastic Labs figures)

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Prime IntellectReinforcement learning primeintellect.ai Open stack for reinforcement learning: compute, environments, evals and deployment. AppAPIOpen sourcePaid

Prime Intellect

Open stack for reinforcement learning: compute, environments, evals and deployment.

Reinforcement learningPaid

What is Prime Intellect?

Prime Intellect sells compute, a reinforcement learning framework, environments, sandboxes and evaluation tools in one stack. Companies use it to train agent models on their own tasks. Its Environments Hub lists community-built RL environments.

What you can do with it

  • Train an agent model on your own workflow with RL
  • Find an RL environment on the Environments Hub
  • Run evaluations and sandboxes

How you use it

AppAPIOpen source

Quick reference

Category
Reinforcement learning
Pricing
Paid
Funding
$130M Series A led by Radical Ventures (Jul 8, 2026)
Valuation (reported)
$1B
Environments Hub (reported)
More than 2,500 environments

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Thinking Machines LabReinforcement learning thinkingmachines.ai Research lab behind Tinker, a managed API for fine-tuning and RL on open models. APIOpen weightsPaid

Thinking Machines Lab

Research lab behind Tinker, a managed API for fine-tuning and RL on open models.

Reinforcement learningPaid

What is Thinking Machines Lab?

Thinking Machines Lab was founded in February 2025 by Mira Murati and others. Tinker, launched in October 2025, lets researchers run fine-tuning and reinforcement learning on open-weight models while the platform runs the GPUs. The lab released the open-weights Inkling models on July 15, 2026.

What you can do with it

  • Run RL experiments on open models without a cluster
  • Write your own training loop in Python
  • Fine-tune Inkling or other open models

How you use it

APIOpen weights

Quick reference

Category
Reinforcement learning
Pricing
Paid
Product
Tinker (Oct 2025)
Open models
Inkling (Jul 15, 2026)
Method
LoRA adapters

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Preference Model (Karotte)Reinforcement learning preferencemodel.com Released Karotte, an RL environment framework, on October 7, 2026. SDKCheck vendor site

Preference Model (Karotte)

Released Karotte, an RL environment framework, on October 7, 2026.

Reinforcement learningCheck vendor site

What is Preference Model?

Preference Model builds tools for reinforcement learning. It released Karotte, a framework to build RL environments, on October 7, 2026.

What you can do with it

  • Build RL environments with a framework
  • Train agents on your own tasks
  • Follow new RL tooling

How you use it

SDK

Quick reference

Category
Reinforcement learning
Pricing
Check vendor site
Release
Karotte (Oct 7, 2026)

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Bespoke LabsReinforcement learning bespokelabs.ai Company-scale RL environments and infrastructure to train reliable agents. APIContact sales

Bespoke Labs

Company-scale RL environments and infrastructure to train reliable agents.

Reinforcement learningContact sales

What is Bespoke Labs?

Bespoke Labs builds reinforcement learning environments and the sandbox infrastructure to run them. It sells environments, not models.

What you can do with it

  • Train agents in company-scale environments
  • Run RL in fast sandboxes
  • Improve agent prompts and policies

How you use it

API

Quick reference

Category
Reinforcement learning
Pricing
Contact sales
Funding
$40M across seed and Series A (Series A led by Wing VC)

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

DeeptuneReinforcement learning deeptune.com RL environments that copy the daily work of accountants, support reps and DevOps engineers. APIContact sales

Deeptune

RL environments that copy the daily work of accountants, support reps and DevOps engineers.

Reinforcement learningContact sales

What is Deeptune?

Deeptune builds high-fidelity reinforcement learning environments. They copy the daily workflows of roles such as accountants, support reps and DevOps engineers, so agents can learn multi-step tasks.

What you can do with it

  • Train agents on realistic job workflows
  • Test agents on multi-step tasks
  • Measure agent progress per role

How you use it

API

Quick reference

Category
Reinforcement learning
Pricing
Contact sales
Funding
$43M Series A led by Andreessen Horowitz (Mar 19, 2026)

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

AfterQueryReinforcement learning afterquery.com Expert-built datasets and RL environments for AI labs. APIContact sales

AfterQuery

Expert-built datasets and RL environments for AI labs.

Reinforcement learningContact sales

What is AfterQuery?

AfterQuery makes datasets and reinforcement learning environments with human experts for AI labs.

What you can do with it

  • Buy expert-built training data
  • Get RL environments for frontier models
  • Evaluate models with expert tasks

How you use it

API

Quick reference

Category
Reinforcement learning
Pricing
Contact sales
Funding
$30M Series A (Apr 2026)

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Andon LabsReinforcement learning andonlabs.com High-fidelity RL environments and evals for frontier coding agents. APIContact sales

Andon Labs

High-fidelity RL environments and evals for frontier coding agents.

Reinforcement learningContact sales

What is Andon Labs?

Andon Labs builds reinforcement learning environments and evaluations for coding agents at frontier labs.

What you can do with it

  • Train coding agents
  • Evaluate agents on realistic tasks
  • Benchmark long tasks

How you use it

API

Quick reference

Category
Reinforcement learning
Pricing
Contact sales
Funding
$9.1M seed (Apr 2026)

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

IdlerReinforcement learning ycombinator.com Frontier data research lab that sells evals, benchmarks and RL environments. APIContact sales

Idler

Frontier data research lab that sells evals, benchmarks and RL environments.

Reinforcement learningContact sales

What is Idler?

Idler builds reinforcement learning environments from real coding work, plus evaluations and benchmarks. It sells them to labs that train frontier models.

What you can do with it

  • Train models on real coding scenarios
  • Benchmark coding agents
  • Buy evals for frontier models

How you use it

API

Quick reference

Category
Reinforcement learning
Pricing
Contact sales
Funding
$9M seed led by Paradigm (Aug 19, 2026)

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

OsmosisReinforcement learning ycombinator.com RL fine-tuning so open models beat foundation models on your task. APIContact sales

Osmosis

RL fine-tuning so open models beat foundation models on your task.

Reinforcement learningContact sales

What is Osmosis?

Osmosis helps companies use reinforcement learning to fine-tune open source models for their own tasks. It is a Y Combinator company (W25).

What you can do with it

  • Fine-tune an open model with RL
  • Lower cost against a large closed model
  • Improve a model on one task

How you use it

API

Quick reference

Category
Reinforcement learning
Pricing
Contact sales
Batch
Y Combinator W25

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

AgileRLReinforcement learning agilerl.com London RLOps platform. Arena is its RL training platform. AppOpen sourceFree tier

AgileRL

London RLOps platform. Arena is its RL training platform.

Reinforcement learningFree tier

What is AgileRL?

AgileRL makes tools to train and tune reinforcement learning agents faster. Arena is its RL training platform.

What you can do with it

  • Train RL agents with less tuning work
  • Use the open source AgileRL library
  • Run RL training on Arena

How you use it

AppOpen source

Quick reference

Category
Reinforcement learning
Pricing
Free tier
Funding
$7.5M seed led by Fusion Fund
Location
London, United Kingdom

Sources

Checked November 10, 2026. Prices and funding change. Check the vendor site before you buy.

Entries checked November 10, 2026. Model labs come from LLM Watch. Companies do not pay to be listed. Do you want to suggest a tool or a correction? Tell us.

Questions about the Directory

Get new entries and research each week. Subscribe.

What is in the A/GEO Research Directory?

The Directory lists the tools and companies that shape AI search visibility. It has AI visibility trackers, AI answer engines, AI crawler control tools and content and schema tools. It also lists the AI companies behind AI answers: model labs, decision models, agent memory and reinforcement learning startups.

How do you pick the entries?

We add a tool or company when it changes how AI answers find, read or cite brands, or when it builds the models and methods behind those answers. Model labs come from LLM Watch, our tracker of AI models. Companies do not pay to be listed.

What is a decision model?

A decision model does not write text. It reads an input and returns one typed answer with a probability: yes or no, one choice from a list, or a score. Agents use decision models to route requests, flag problems and check work, fast and at low cost.

Why are reinforcement learning startups in an AI search directory?

Reinforcement learning is how AI labs train models to use tools, search the web and finish tasks. The environments these startups build shape how AI agents search and which sources they trust. Agents are a new kind of visitor to your website.

How current are the prices and funding numbers?

Each entry shows the date we checked it and links to its sources. Prices and funding change often. Check the vendor site before you buy.

How do I suggest a tool or a correction?

Use the contact form and tell us the tool, its website and what it does. For a correction, give the entry name and a source for the right fact.

Find out what AI says about you.

We ask leading AI answer engines the questions your customers ask. Then we show you who they name, and what to fix first. The check is free.

Get a free AI visibility check

We ask leading AI answer engines the questions your customers ask. Then we send you a short report.

We use your details only to send your report and reply to you. See our privacy policy.