Top pick
ChatGPT Plus
$20/mo
GPT-6.2 Sol for everyday work and GPT-6.2 Astra for hard problems, plus image generation, voice and deep research.
See plan detailsPrices checked
25 models from 14 AI labs in the US, China and Europe, plus the 8 most-used AI apps. Prices, context windows, plans and features side by side.
US dollars per million tokens through each company’s API. Log scale, so every step right is about 10× pricier.
Choose what you mostly want to do. You’ll get three apps that fit, what they cost each month, and why.
Top pick
$20/mo
GPT-6.2 Sol for everyday work and GPT-6.2 Astra for hard problems, plus image generation, voice and deep research.
See plan detailsAlso great
$20/mo
Strong long-form writing and analysis of big documents, with Projects to keep your files together. $17/mo billed yearly.
See plan detailsWorth a look
$19.99/mo
Gemini 4 Argon plus 5 TB of Google storage, inside Gmail, Docs and Android.
See plan detailsPicks are our own editorial choices, not paid placements.
Subscription prices for people using the chat apps, not the API. Most companies now sell a $5–$20 tier for everyday use and a $100–$200 tier for heavy use.
OpenAI
Free and Go use GPT-6 Luna and may show ads. Plus includes GPT-6.2 Sol and GPT-6.2 Astra; Pro raises limits and adds Pro reasoning.
Anthropic
Pro is $17/mo billed yearly. Every paid plan includes Claude Code. Max and Ultra include Fable; on Pro it uses usage credits.
Google AI plans
AI Plus runs Gemini 4 Hydrogen, Pro adds Gemini 4 Argon, Ultra adds Gemini 4 Carbon. Plans bundle 400 GB to 30 TB of storage.
Sold as the X Subscription
X Premium adds Cursor Pro, and every paid tier includes premium features on X. Paid tiers share a weekly Grok usage pool.
Perplexity
Pro lets you choose models from OpenAI, Anthropic, Google and xAI. Max adds their top models and the most Perplexity Computer credits.
Mistral (formerly Le Chat)
Mistral models only. Students get Pro for $5.99.
Meta, with Muse Code for coding
Muse Wonder is included from $5. Power and Maximum add Muse Wonder Full, Meta’s top model.
DeepSeek
No subscription. The app is free; developers pay per token.
Built-in features in each company’s own app. Some need a paid plan. For models you can download and run, see open-source AI models.
YesLimited or partly availableNot offered or not confirmed
| App | Web search | Image generation | Video generation | Voice chat | Agent that acts for you | Coding agent | Deep research |
|---|---|---|---|---|---|---|---|
| ChatGPT | |||||||
| Claude | |||||||
| Gemini | |||||||
| Grok | |||||||
| Meta AI | |||||||
| Perplexity | |||||||
| Mistral Vibe | |||||||
| DeepSeek |
The current flagship and budget models each company sells through its API. Sort any column, or filter by where the lab is based or by company. Context is how much text the model can read at once; 1M tokens is roughly 750,000 words.
Showing all 25 models, cheapest first.
| Best for | |||||
|---|---|---|---|---|---|
GPT-6 Luna OpenAI, US |
Sep 22, 2026 | 1.05M | $0.10 | $0.50 | High-volume tasks; runs ChatGPT Free and Go |
Claude Haiku 5.5 Anthropic, US |
Oct 7, 2026 | 1M | $0.10 | $0.50 | Fast, cheap, high volume |
Hy3 Tencent, China |
Jul 6, 2026 | 256K | $0.132 | $0.528 | Cheapest open-weight model here (Apache-2.0) |
MiMo-v2.6 Pro Xiaomi, China |
Oct 2026 | Not stated | $0.435 | $0.87 | Very low-cost reasoning model from Xiaomi |
DeepSeek V4.1 Flash DeepSeek, China |
Sep 10, 2026 | 1M | $0.30 | $1.20 | Fast open model; half price off-peak ($0.15 / $0.60) |
MiniMax M3 MiniMax, China |
Jun 1, 2026 | 1M | $0.30 | $1.20 | Low-cost model that reads images and video; community license |
Mistral Large 4 Mistral, Europe |
Oct 6, 2026 | 1M | $0.70 | $2.10 | Mistral’s flagship, with open weights under Mistral’s license |
Doubao Seed 2.0 Pro ByteDance, China |
Feb 2026 | 256K | $0.50 | $3.00 | Model behind China’s most-used chatbot, Doubao |
DeepSeek V4 Pro DeepSeek, China |
Aug 13, 2026 | 1M | $1.32 | $3.96 | Strong open model; half price off-peak ($0.66 / $1.98) |
Muse Spark 1.3 Meta, US |
Sep 2, 2026 | 1M | $1.25 | $4.25 | Multimodal reasoning at a lower price |
GLM-5.3 Zhipu, China |
Aug 2026 | 1M | $1.40 | $4.40 | Zhipu’s newest flagship, strong at coding |
Qwen4 Max Alibaba, China |
Nov 2026 | 1M | $1.50 | $4.50 | Alibaba’s newest flagship, now on pay-as-you-go |
GPT-6.1 Terra OpenAI, US |
Oct 26, 2026 | 1.05M | $1.00 | $5.00 | Fast, efficient everyday work |
Grok 4.7 xAI, US |
Sep 21, 2026 | 500K | $2.00 | $6.00 | Coding, agents and knowledge work |
Gemini 4 Hydrogen Google, US |
Nov 12, 2026 | 1.05M | $1.50 | $7.50 | Everyday coding and agent workflows; free tier |
Mistral Medium 3.5 Mistral, Europe |
Apr 28, 2026 | 256K | $1.50 | $7.50 | Being retired in April 2027; Large 4 replaces it |
GPT-6.2 Sol OpenAI, US |
Nov 17, 2026 | 1.05M | $2.00 | $10.00 | Close to Astra quality at a quarter of the price |
Claude Sonnet 5.5 Anthropic, US |
Sep 28, 2026 | 1M | $2.00 | $10.00 | Balanced quality and cost |
Gemini 4 Carbon Google, US |
Nov 10, 2026 | 1.05M | $2.00 | $10.00 | Google’s model for the hardest reasoning problems |
Gemini 4 Argon Google, US |
Oct 13, 2026 | 1.05M | $2.00 | $10.00 | Deep reasoning with faster replies than Carbon |
Kimi K3.1 Pro Moonshot, China |
Oct 23, 2026 | 1M | $2.00 | $10.00 | Coding and agents; weights released Nov 5 |
Claude Opus 5.5 Anthropic, US |
Sep 22, 2026 | 1M | $4.00 | $20.00 | Long, complex tasks and agents |
Muse Wonder Meta, US |
Oct 12, 2026 | 1M | $4.00 | $20.00 | Meta’s frontier model for agents |
GPT-6.2 Astra OpenAI, US |
Nov 17, 2026 | 1.05M | $8.00 | $40.00 | OpenAI’s most capable model, for the hardest reasoning |
Claude Fable 5.2 Anthropic, US |
Nov 9, 2026 | 1M | $8.00 | $40.00 | Anthropic’s top model for coding and knowledge work |
Head-to-head pages, pricing guides and tools for the questions people search for most, built from the same checked prices.
Short answers about how this comparison works and how to read it.
For most people, a $20 plan: ChatGPT Plus for the broadest set of tools, Claude Pro if you write or code a lot, or Google AI Pro ($19.99) if you also want video and 5 TB of storage. Use the picker above or the ChatGPT vs Claude comparison to decide.
GPT-6 Luna and Claude Haiku 5.5, at $0.10 input and $0.50 output per million tokens. The cheapest open-weight model is Tencent Hy3 at $0.132 / $0.528. See every model ranked on the cheapest AI API page.
The app is the chat you use at chatgpt.com, claude.ai or gemini.google.com, billed as a monthly plan. The API is how developers build the models into their own products, billed per token used. A model can be cheap on the API and still need a paid plan in the app.
A token is a chunk of text, about three-quarters of a word in English. Input tokens are what you send the model; output tokens are what it writes back. Output usually costs two to five times more.
It depends on the job. All of the flagship models now handle writing, coding and analysis well. The bigger differences are price, the tools built around them (coding agents, image and video tools, research modes), and how much you can use them before hitting a limit.
An open-weight model can be downloaded and run on your own hardware. DeepSeek, Moonshot’s Kimi, Zhipu’s GLM, MiniMax, Tencent and Mistral publish weights for the models marked “Open weights” here, though some licenses limit commercial use. See open-source AI models. It suits companies that need full control of their data or want to avoid per-token fees.
Prices and models change almost monthly. Everything here was checked on November 18, 2026 against company pricing pages and release notes. Always confirm on the company’s own site before you buy.
No. This site isn’t affiliated with OpenAI, Anthropic, Google, xAI, Meta, DeepSeek, Mistral or Perplexity. Product names are trademarks of their owners.