<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: AI-Sovereign</title>
    <description>The latest articles on DEV Community by AI-Sovereign (@ai_sovereign).</description>
    <link>https://dev.to/ai_sovereign</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4258376%2F3d80aed5-830e-45d2-b88b-d9a4de1a75bb.png</url>
      <title>DEV Community: AI-Sovereign</title>
      <link>https://dev.to/ai_sovereign</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/ai_sovereign"/>
    <language>en</language>
    <item>
      <title>The 20 biggest open-source AI agent projects on GitHub (November 2026)</title>
      <dc:creator>AI-Sovereign</dc:creator>
      <pubDate>Sun, 22 Nov 2026 07:18:14 +0000</pubDate>
      <link>https://dev.to/ai_sovereign/the-20-biggest-open-source-ai-agent-projects-on-github-november-2026-5dpg</link>
      <guid>https://dev.to/ai_sovereign/the-20-biggest-open-source-ai-agent-projects-on-github-november-2026-5dpg</guid>
      <description>&lt;p&gt;AI agents are the fastest-moving corner of open source. We track &lt;strong&gt;576 agent projects&lt;/strong&gt; on GitHub with more than 500 stars in the &lt;a href="https://ai-sovereign.com/agents" rel="noopener noreferrer"&gt;AI-Sovereign directory&lt;/a&gt;, and refresh their numbers every day. Here are the 20 biggest on 22 November 2026, ranked by stars, with what each one actually does.&lt;/p&gt;

&lt;h2&gt;
  
  
  The top 20
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;#&lt;/th&gt;
&lt;th&gt;Project&lt;/th&gt;
&lt;th&gt;Stars&lt;/th&gt;
&lt;th&gt;Language&lt;/th&gt;
&lt;th&gt;What it does&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;1&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/affaan-m-ecc" rel="noopener noreferrer"&gt;ECC&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;308K&lt;/td&gt;
&lt;td&gt;JavaScript&lt;/td&gt;
&lt;td&gt;Performance system for agent harnesses: skills, instincts, memory, security&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;2&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/deepseek-ai-deepseek-harness" rel="noopener noreferrer"&gt;DeepSeek Harness&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;285K&lt;/td&gt;
&lt;td&gt;TypeScript&lt;/td&gt;
&lt;td&gt;DeepSeek's agent harness, built so everything is a plugin&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;3&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/nousresearch-hermes-agent" rel="noopener noreferrer"&gt;Hermes Agent&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;265K&lt;/td&gt;
&lt;td&gt;Python&lt;/td&gt;
&lt;td&gt;Nous Research's personal agent that grows with its user&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;4&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/firecrawl-firecrawl" rel="noopener noreferrer"&gt;Firecrawl&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;206K&lt;/td&gt;
&lt;td&gt;TypeScript&lt;/td&gt;
&lt;td&gt;Search, scraping and data access so agents can read the web&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;5&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/dietrichgebert-ponytail" rel="noopener noreferrer"&gt;Ponytail&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;196K&lt;/td&gt;
&lt;td&gt;JavaScript&lt;/td&gt;
&lt;td&gt;Makes coding agents write less code ("the best code is the code you never wrote")&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;6&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/graphify-labs-graphify" rel="noopener noreferrer"&gt;Graphify&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;152K&lt;/td&gt;
&lt;td&gt;Python&lt;/td&gt;
&lt;td&gt;Turns a codebase, docs and schemas into a queryable knowledge graph&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;7&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/langchain-ai-langchain" rel="noopener noreferrer"&gt;LangChain&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;150K&lt;/td&gt;
&lt;td&gt;Python&lt;/td&gt;
&lt;td&gt;The agent engineering platform&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;8&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/stablyai-orca" rel="noopener noreferrer"&gt;Orca&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;132K&lt;/td&gt;
&lt;td&gt;TypeScript&lt;/td&gt;
&lt;td&gt;Runs a fleet of coding bots on your own subscriptions&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;9&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/browser-use-browser-use" rel="noopener noreferrer"&gt;Browser Use&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;122K&lt;/td&gt;
&lt;td&gt;Python&lt;/td&gt;
&lt;td&gt;Library for AI agents that drive a real browser&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;10&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/leonxlnx-taste-skill" rel="noopener noreferrer"&gt;Taste Skill&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;120K&lt;/td&gt;
&lt;td&gt;JavaScript&lt;/td&gt;
&lt;td&gt;Gives coding agents design taste: art direction, references, a browser check&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;11&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/thedotmack-claude-mem" rel="noopener noreferrer"&gt;claude-mem&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;119K&lt;/td&gt;
&lt;td&gt;TypeScript&lt;/td&gt;
&lt;td&gt;Persistent memory across agent sessions&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;12&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/paperclipai-paperclip" rel="noopener noreferrer"&gt;Paperclip&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;116K&lt;/td&gt;
&lt;td&gt;TypeScript&lt;/td&gt;
&lt;td&gt;Open-source app for managing agents at work&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;13&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/panniantong-agent-reach" rel="noopener noreferrer"&gt;Agent Reach&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;114K&lt;/td&gt;
&lt;td&gt;Python&lt;/td&gt;
&lt;td&gt;Lets agents read and search social platforms and video sites&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;14&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/google-gemini-gemini-cli" rel="noopener noreferrer"&gt;Gemini CLI&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;109K&lt;/td&gt;
&lt;td&gt;TypeScript&lt;/td&gt;
&lt;td&gt;Google's open-source terminal agent&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;15&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/tt-a1i-archify" rel="noopener noreferrer"&gt;Archify&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;99K&lt;/td&gt;
&lt;td&gt;JavaScript&lt;/td&gt;
&lt;td&gt;Agent skill that turns ideas and code into interactive diagrams&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;16&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/infiniflow-ragflow" rel="noopener noreferrer"&gt;RAGFlow&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;95K&lt;/td&gt;
&lt;td&gt;Go&lt;/td&gt;
&lt;td&gt;Retrieval engine that grew into context infrastructure for agents&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;17&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/cheaper-inference-omniroute" rel="noopener noreferrer"&gt;OmniRoute&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;89K&lt;/td&gt;
&lt;td&gt;TypeScript&lt;/td&gt;
&lt;td&gt;Free AI gateway: one endpoint for many model providers&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;18&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/heygen-com-hyperframes" rel="noopener noreferrer"&gt;HyperFrames&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;88K&lt;/td&gt;
&lt;td&gt;TypeScript&lt;/td&gt;
&lt;td&gt;Open-source AI video editor driven by text&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;19&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/unclecode-crawl4ai" rel="noopener noreferrer"&gt;Crawl4AI&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;88K&lt;/td&gt;
&lt;td&gt;Python&lt;/td&gt;
&lt;td&gt;Web crawling and extraction built for agents&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;20&lt;/td&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/agent/bytedance-deer-flow" rel="noopener noreferrer"&gt;DeerFlow&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;86K&lt;/td&gt;
&lt;td&gt;Python&lt;/td&gt;
&lt;td&gt;ByteDance's long-horizon agent that researches, codes and creates&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  What the list says
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;The harness is the product now.&lt;/strong&gt; All of the top three are not models or apps but harnesses: the layer that wraps a model with tools, memory, skills and rules (&lt;a href="https://ai-sovereign.com/p/agent/affaan-m-ecc" rel="noopener noreferrer"&gt;ECC&lt;/a&gt;, &lt;a href="https://ai-sovereign.com/p/agent/deepseek-ai-deepseek-harness" rel="noopener noreferrer"&gt;DeepSeek Harness&lt;/a&gt;, &lt;a href="https://ai-sovereign.com/p/agent/nousresearch-hermes-agent" rel="noopener noreferrer"&gt;Hermes Agent&lt;/a&gt;). The model is becoming a replaceable part; the harness decides how good the agent feels.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Skills are a category of their own.&lt;/strong&gt; &lt;a href="https://ai-sovereign.com/p/agent/dietrichgebert-ponytail" rel="noopener noreferrer"&gt;Ponytail&lt;/a&gt;, &lt;a href="https://ai-sovereign.com/p/agent/leonxlnx-taste-skill" rel="noopener noreferrer"&gt;Taste Skill&lt;/a&gt; and &lt;a href="https://ai-sovereign.com/p/agent/tt-a1i-archify" rel="noopener noreferrer"&gt;Archify&lt;/a&gt; are not agents at all. They are packaged know-how that you drop into an existing coding agent to change how it works. Three of them sit in the top 15.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Agents need eyes and memory.&lt;/strong&gt; Five projects in the top 20 exist to give agents access to the outside world or a memory of the past: &lt;a href="https://ai-sovereign.com/p/agent/firecrawl-firecrawl" rel="noopener noreferrer"&gt;Firecrawl&lt;/a&gt;, &lt;a href="https://ai-sovereign.com/p/agent/unclecode-crawl4ai" rel="noopener noreferrer"&gt;Crawl4AI&lt;/a&gt; and &lt;a href="https://ai-sovereign.com/p/agent/panniantong-agent-reach" rel="noopener noreferrer"&gt;Agent Reach&lt;/a&gt; for web data, &lt;a href="https://ai-sovereign.com/p/agent/thedotmack-claude-mem" rel="noopener noreferrer"&gt;claude-mem&lt;/a&gt; and &lt;a href="https://ai-sovereign.com/p/agent/graphify-labs-graphify" rel="noopener noreferrer"&gt;Graphify&lt;/a&gt; for memory and knowledge. A model alone does not know what happened today or what you told it last week.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Python and TypeScript split the field.&lt;/strong&gt; Across all 576 agent projects we track, 182 are written mainly in Python and 177 in TypeScript, far ahead of JavaScript (45), Rust (39) and Go (35).&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Almost everything ships under a permissive license.&lt;/strong&gt; 19 of the top 20 use MIT or Apache 2.0; &lt;a href="https://ai-sovereign.com/p/agent/firecrawl-firecrawl" rel="noopener noreferrer"&gt;Firecrawl&lt;/a&gt; is the notable AGPL exception.&lt;/p&gt;

&lt;h2&gt;
  
  
  Methodology
&lt;/h2&gt;

&lt;p&gt;We collect GitHub repositories tagged with agent topics that have more than 500 stars, rank them by stars and refresh the numbers every day. Each project page shows its stars, forks, language, license, last commit and its 24-hour and 7-day change once history builds up. See the live ranking on the &lt;a href="https://ai-sovereign.com/agents" rel="noopener noreferrer"&gt;AI Agents page&lt;/a&gt; and the wider &lt;a href="https://ai-sovereign.com/open-source" rel="noopener noreferrer"&gt;open-source AI ranking&lt;/a&gt;, and read &lt;a href="https://ai-sovereign.com/about" rel="noopener noreferrer"&gt;how AI-Sovereign ranks projects&lt;/a&gt;.&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://ai-sovereign.com/blog/top-ai-agent-projects-github-november-2026" rel="noopener noreferrer"&gt;AI-Sovereign&lt;/a&gt;, the daily ranking of AI models, APIs, agents and open-source projects.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>agents</category>
      <category>opensource</category>
      <category>github</category>
    </item>
    <item>
      <title>AI model API prices in November 2026: from $0.02 to $180 per million tokens</title>
      <dc:creator>AI-Sovereign</dc:creator>
      <pubDate>Sun, 22 Nov 2026 07:18:07 +0000</pubDate>
      <link>https://dev.to/ai_sovereign/ai-model-api-prices-in-november-2026-from-002-to-180-per-million-tokens-3g6</link>
      <guid>https://dev.to/ai_sovereign/ai-model-api-prices-in-november-2026-from-002-to-180-per-million-tokens-3g6</guid>
      <description>&lt;p&gt;Picking an AI model used to be a question of quality. In late 2026 it is just as much a question of price, because the gap between the cheapest and the most expensive model you can call by API is now enormous.&lt;/p&gt;

&lt;p&gt;We track &lt;strong&gt;373 AI model APIs&lt;/strong&gt; in the &lt;a href="https://ai-sovereign.com/apis" rel="noopener noreferrer"&gt;AI-Sovereign directory&lt;/a&gt;, with input and output prices per million tokens and the context size of each one, refreshed every day. This report looks at what the numbers say on 22 November 2026.&lt;/p&gt;

&lt;h2&gt;
  
  
  The short version
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;The cheapest capable models cost &lt;strong&gt;$0.02 to $0.10 per million input tokens&lt;/strong&gt;. At that price, a million short chatbot replies cost a few dollars.&lt;/li&gt;
&lt;li&gt;The most expensive model on the list, &lt;strong&gt;GPT-5.5 Pro&lt;/strong&gt;, costs &lt;strong&gt;$30 in and $180 out&lt;/strong&gt; per million tokens. That is 6,000 times the output price of the cheapest model on the list ($0.03 for Mistral Nemo).&lt;/li&gt;
&lt;li&gt;A &lt;strong&gt;one-million-token context window&lt;/strong&gt; is no longer rare. OpenAI, Google, Qwen, DeepSeek, Z.ai and NVIDIA all sell models that read about a million tokens at once, and two Grok models read two million.&lt;/li&gt;
&lt;li&gt;Only &lt;strong&gt;3&lt;/strong&gt; of the 373 APIs are listed as free.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  The cheapest models from the big labs
&lt;/h2&gt;

&lt;p&gt;These are the lowest prices we see from established labs, for models with at least 128K tokens of context:&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Model&lt;/th&gt;
&lt;th&gt;Lab&lt;/th&gt;
&lt;th&gt;Input / 1M tokens&lt;/th&gt;
&lt;th&gt;Output / 1M tokens&lt;/th&gt;
&lt;th&gt;Context&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/meta-llama-llama-3-1-8b-instruct" rel="noopener noreferrer"&gt;Llama 3.1 8B Instruct&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;Meta&lt;/td&gt;
&lt;td&gt;$0.02&lt;/td&gt;
&lt;td&gt;$0.04&lt;/td&gt;
&lt;td&gt;131K&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/mistralai-mistral-nemo" rel="noopener noreferrer"&gt;Mistral Nemo&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;Mistral&lt;/td&gt;
&lt;td&gt;$0.029&lt;/td&gt;
&lt;td&gt;$0.03&lt;/td&gt;
&lt;td&gt;131K&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/google-gemma-3-4b-it" rel="noopener noreferrer"&gt;Gemma 3 4B&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;Google&lt;/td&gt;
&lt;td&gt;$0.04&lt;/td&gt;
&lt;td&gt;$0.05&lt;/td&gt;
&lt;td&gt;131K&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/openai-gpt-oss-20b" rel="noopener noreferrer"&gt;gpt-oss-20b&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;OpenAI&lt;/td&gt;
&lt;td&gt;$0.017&lt;/td&gt;
&lt;td&gt;$0.085&lt;/td&gt;
&lt;td&gt;131K&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/z-ai-glm-5-3-flash" rel="noopener noreferrer"&gt;GLM 5.3 Flash&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;Z.ai&lt;/td&gt;
&lt;td&gt;$0.04&lt;/td&gt;
&lt;td&gt;$0.078&lt;/td&gt;
&lt;td&gt;1M&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/qwen-qwen3-7-flash" rel="noopener noreferrer"&gt;Qwen3.7 Flash&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;Qwen&lt;/td&gt;
&lt;td&gt;$0.03&lt;/td&gt;
&lt;td&gt;$0.13&lt;/td&gt;
&lt;td&gt;1M&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/deepseek-deepseek-v4-1-flash" rel="noopener noreferrer"&gt;DeepSeek V4.1 Flash&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;DeepSeek&lt;/td&gt;
&lt;td&gt;$0.05&lt;/td&gt;
&lt;td&gt;$0.20&lt;/td&gt;
&lt;td&gt;1M&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/nvidia-nemotron-3-5-lightning" rel="noopener noreferrer"&gt;Nemotron 3.5 Lightning&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;NVIDIA&lt;/td&gt;
&lt;td&gt;$0.07&lt;/td&gt;
&lt;td&gt;$0.20&lt;/td&gt;
&lt;td&gt;1M&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Two things stand out. First, the cheapest end of the market is dominated by open-weight models (Llama, Mistral, Gemma, gpt-oss), which many providers compete to serve. Second, the "flash" tier now comes with a million-token context: &lt;a href="https://ai-sovereign.com/p/api/z-ai-glm-5-3-flash" rel="noopener noreferrer"&gt;GLM 5.3 Flash&lt;/a&gt;, &lt;a href="https://ai-sovereign.com/p/api/qwen-qwen3-7-flash" rel="noopener noreferrer"&gt;Qwen3.7 Flash&lt;/a&gt; and &lt;a href="https://ai-sovereign.com/p/api/deepseek-deepseek-v4-1-flash" rel="noopener noreferrer"&gt;DeepSeek V4.1 Flash&lt;/a&gt; read a whole codebase or a stack of contracts for a few cents.&lt;/p&gt;

&lt;h2&gt;
  
  
  The most expensive models
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Model&lt;/th&gt;
&lt;th&gt;Lab&lt;/th&gt;
&lt;th&gt;Input / 1M tokens&lt;/th&gt;
&lt;th&gt;Output / 1M tokens&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/openai-gpt-5-5-pro" rel="noopener noreferrer"&gt;GPT-5.5 Pro&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;OpenAI&lt;/td&gt;
&lt;td&gt;$30&lt;/td&gt;
&lt;td&gt;$180&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;GPT-5.4 Pro&lt;/td&gt;
&lt;td&gt;OpenAI&lt;/td&gt;
&lt;td&gt;$30&lt;/td&gt;
&lt;td&gt;$180&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;GPT-5.2 Pro&lt;/td&gt;
&lt;td&gt;OpenAI&lt;/td&gt;
&lt;td&gt;$21&lt;/td&gt;
&lt;td&gt;$168&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;GPT-5 Pro&lt;/td&gt;
&lt;td&gt;OpenAI&lt;/td&gt;
&lt;td&gt;$15&lt;/td&gt;
&lt;td&gt;$120&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;o3 Pro&lt;/td&gt;
&lt;td&gt;OpenAI&lt;/td&gt;
&lt;td&gt;$20&lt;/td&gt;
&lt;td&gt;$80&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/anthropic-claude-opus-4-1" rel="noopener noreferrer"&gt;Claude Opus 4.1&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;Anthropic&lt;/td&gt;
&lt;td&gt;$15&lt;/td&gt;
&lt;td&gt;$75&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;The top of the price list belongs to the "Pro" reasoning models, which spend far more compute on every answer. They make sense for hard, high-value questions where one good answer is worth dollars. For anything that runs at volume, they are the wrong tool.&lt;/p&gt;

&lt;h2&gt;
  
  
  What the newest flagships cost
&lt;/h2&gt;

&lt;p&gt;The newest frontier models from the big labs cluster in two price bands:&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Model&lt;/th&gt;
&lt;th&gt;Lab&lt;/th&gt;
&lt;th&gt;Input / 1M&lt;/th&gt;
&lt;th&gt;Output / 1M&lt;/th&gt;
&lt;th&gt;Context&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/openai-gpt-6-2-astra" rel="noopener noreferrer"&gt;GPT-6.2 Astra&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;OpenAI&lt;/td&gt;
&lt;td&gt;$8&lt;/td&gt;
&lt;td&gt;$40&lt;/td&gt;
&lt;td&gt;1.05M&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/anthropic-claude-fable-5-2" rel="noopener noreferrer"&gt;Claude Fable 5.2&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;Anthropic&lt;/td&gt;
&lt;td&gt;$8&lt;/td&gt;
&lt;td&gt;$40&lt;/td&gt;
&lt;td&gt;1M&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/google-gemini-4-neon" rel="noopener noreferrer"&gt;Gemini 4 Neon&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;Google&lt;/td&gt;
&lt;td&gt;$3&lt;/td&gt;
&lt;td&gt;$15&lt;/td&gt;
&lt;td&gt;1M&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/openai-gpt-6-2-sol" rel="noopener noreferrer"&gt;GPT-6.2 Sol&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;OpenAI&lt;/td&gt;
&lt;td&gt;$2&lt;/td&gt;
&lt;td&gt;$10&lt;/td&gt;
&lt;td&gt;1.05M&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/google-gemini-4-hydrogen" rel="noopener noreferrer"&gt;Gemini 4 Hydrogen&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;Google&lt;/td&gt;
&lt;td&gt;$1.50&lt;/td&gt;
&lt;td&gt;$7.50&lt;/td&gt;
&lt;td&gt;1M&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/qwen-qwen4-max" rel="noopener noreferrer"&gt;Qwen4 Max&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;Qwen&lt;/td&gt;
&lt;td&gt;$1.43&lt;/td&gt;
&lt;td&gt;$4.28&lt;/td&gt;
&lt;td&gt;1M&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/openai-gpt-6-1-terra" rel="noopener noreferrer"&gt;GPT-6.1 Terra&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;OpenAI&lt;/td&gt;
&lt;td&gt;$1&lt;/td&gt;
&lt;td&gt;$5&lt;/td&gt;
&lt;td&gt;1.05M&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;The top flagships sit at $8 in and $40 out. One step down, at $1 to $3 in, every lab now sells a model with a million-token context, which is where most production workloads will land.&lt;/p&gt;

&lt;h2&gt;
  
  
  The biggest context windows
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Model&lt;/th&gt;
&lt;th&gt;Lab&lt;/th&gt;
&lt;th&gt;Context&lt;/th&gt;
&lt;th&gt;Input / 1M&lt;/th&gt;
&lt;th&gt;Output / 1M&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/x-ai-grok-4-20" rel="noopener noreferrer"&gt;Grok 4.20&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;xAI&lt;/td&gt;
&lt;td&gt;2M&lt;/td&gt;
&lt;td&gt;$1.25&lt;/td&gt;
&lt;td&gt;$2.50&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/meta-llama-llama-4-scout" rel="noopener noreferrer"&gt;Llama 4 Scout&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;Meta&lt;/td&gt;
&lt;td&gt;1.31M&lt;/td&gt;
&lt;td&gt;$0.10&lt;/td&gt;
&lt;td&gt;$0.30&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;a href="https://ai-sovereign.com/p/api/openai-gpt-6-2-astra" rel="noopener noreferrer"&gt;GPT-6.2 Astra&lt;/a&gt;&lt;/td&gt;
&lt;td&gt;OpenAI&lt;/td&gt;
&lt;td&gt;1.05M&lt;/td&gt;
&lt;td&gt;$8&lt;/td&gt;
&lt;td&gt;$40&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;&lt;a href="https://ai-sovereign.com/p/api/meta-llama-llama-4-scout" rel="noopener noreferrer"&gt;Llama 4 Scout&lt;/a&gt; is the surprise here: 1.3 million tokens of context for $0.10 in and $0.30 out.&lt;/p&gt;

&lt;h2&gt;
  
  
  How to read these numbers
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Price is per million tokens.&lt;/strong&gt; A token is roughly three quarters of an English word, so a million tokens is about 750,000 words.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Output usually costs more than input,&lt;/strong&gt; often four to six times more. A chatbot that writes long answers pays mostly for output; a tool that reads long documents and returns a short summary pays mostly for input.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;The cheapest model is rarely the cheapest system.&lt;/strong&gt; A model that needs three attempts costs more than a better one that gets it right the first time. Test your real prompts on two or three price tiers before you commit.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Methodology
&lt;/h2&gt;

&lt;p&gt;Prices and context sizes come from OpenRouter's public model list, which shows the price of each model across the providers that serve it. We refresh them every day. See every model, sorted by price or context, on the &lt;a href="https://ai-sovereign.com/apis?sort=price" rel="noopener noreferrer"&gt;AI Model APIs page&lt;/a&gt;, and read &lt;a href="https://ai-sovereign.com/about" rel="noopener noreferrer"&gt;how AI-Sovereign ranks projects&lt;/a&gt;.&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://ai-sovereign.com/blog/ai-model-api-prices-november-2026" rel="noopener noreferrer"&gt;AI-Sovereign&lt;/a&gt;, the daily ranking of AI models, APIs, agents and open-source projects.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>llm</category>
      <category>api</category>
      <category>pricing</category>
    </item>
  </channel>
</rss>
