Free site where you compare AI models side by side and see which ones people rate best.

Artificial Analysis
Free, independent scores for how smart, fast and cheap each AI model is.
Quick take
- A free, independent scoreboard for AI models: one Intelligence Index score, plus speed and cost for each model.
- Covers more than 500 models across text, image, video, speech, music and coding agents.
- Image, video and speech rankings come from blind votes, where people pick the better result without knowing the model.
- It's a reference site for choosing models. Test the top picks on your own prompts before you commit.
What is Artificial Analysis?
Artificial Analysis is an independent company that tests AI models and publishes the results for free. Its Intelligence Index gives each language model one score based on a set of hard tests, next to how fast it writes and what a typical task costs. It covers more than 500 models, from ChatGPT and Claude to open models you can run yourself, and the companies that serve them.
It also ranks image, video, speech and music models, using blind votes where people pick the better result without knowing which model made it, and it tests coding agents too. The Model Recommender suggests a model based on whether you care most about quality, speed or price. AI labs, journalists and investors quote its charts all the time, and developers can pull the data through a free API.
The catch is that it's a reference site. You check it before choosing, then do your work elsewhere. It tells you which model scores best on its tests, which isn't always the one that writes best for your business. Use it to shortlist, then try the top two or three on your own prompts.
Who's using it, and who it's for
- Developers
- Founders
- Journalists
- AI enthusiasts
- Investors
Anyone choosing an AI model or tool who wants independent numbers on quality, speed and price before paying.
Key features
Intelligence Index
One score per language model, built from a set of independent tests, so you can compare models at a glance.
Speed and cost
How fast each model writes and what an average task costs, so you can weigh quality against price.
Media arenas
Leaderboards for image, video, speech and music models based on blind votes.
Coding agent index
Scores for coding agents, the model plus its tools, on real multi-step tasks.
Model Recommender
Tell it what matters most to you and it suggests a model.
Free API
Pull benchmark data into your own tools or ask Claude to fetch it.
How we rate it
4.3 / 5
Our rating
From using the tool ourselves. The average of the five scores.
- Ease of use
- 3.5
- Features
- 4.0
- Output quality
- 4.5
- Value for money
- 5.0
- Speed & reliability
- 4.5
How quickly someone non-technical gets value from it.
How completely it covers its own job, and how well the pieces fit together.
How good the results are before you fix them.
What you get for the price, including the free plan.
How fast it is and how often it falls over.
Artificial Analysis pricing
- Our pick
Free
Free
Everyone choosing between AI models.
- All public leaderboards and charts
- Model Recommender
- Free API with attribution
Commercial
Custom
Companies that need deeper data or custom testing.
- Commercial API with more detailed data
- Custom evaluations
Which plan should you pick?
Use the free site. The paid data and custom testing are for AI companies and large businesses comparing models at scale.
Prices in USD unless noted, checked 3 November 2026. Vendors change plans often, so confirm on their site before you buy.
Best use cases
Ways Artificial Analysis can slot into your week.
- 1
Check a new model on launch day
When a new model is announced, see where it lands on the Intelligence Index before you switch.
- 2
Choose the cheapest model that's good enough
Compare price per task against score to pick a model for a chatbot or automation.
- 3
Pick an image or voice model
Use the image and text-to-speech leaderboards to shortlist tools before you sign up.
How it fits your tech stack
Use it alongside LMArena when choosing models, then OpenRouter or the model makers' own apps to try them. It doesn't replace testing on your own work.
Pros and cons
Pros
- Free and independent
- Covers text, image, video, speech and coding agents
- Shows speed and price, not just quality
- Updated quickly when new models launch
- Free API for the data
Cons
- A reference site you check, with nothing to do your work in
- Test scores don't always match how a model handles your tasks
- Dense charts take some getting used to
Artificial Analysis alternatives
The closest alternatives to Artificial Analysis in the library, best rated first.

Claude
4.6
Best for long-form writing and editing
Compare with Artificial Analysis
ChatGPT
4.3
Best for everyday questions and brainstorming
Compare with Artificial AnalysisBest for searching and summarising your Gmail
Compare with Artificial Analysis
Langdock
4.2
Best for rolling out AI to a whole company in the EU
Compare with Artificial Analysis
Why it made the list
It's the most quoted independent benchmark for AI models, it's free, and it covers text, image, video, speech and coding agents in one place.
Artificial Analysis FAQ
How much does Artificial Analysis cost?
It's free to use. Prices checked 3 November 2026 against Artificial Analysis's pricing page.
Is Artificial Analysis free?
Yes, Artificial Analysis has a free option. Use the free site. The paid data and custom testing are for AI companies and large businesses comparing models at scale.
What is Artificial Analysis best for?
Artificial Analysis is best for checking which AI model is best right now, Comparing API prices and speeds and picking an image, video or voice model. Anyone choosing an AI model or tool who wants independent numbers on quality, speed and price before paying.
What are the best Artificial Analysis alternatives?
The closest alternatives are Claude, ChatGPT, Google Gemini and Langdock. Claude is best for long-form writing and editing.
More tools in AI Assistants & LLMs


One account and API key for hundreds of AI models, at the makers' prices.

The home of open AI models: download, try and run millions of them in one place.

OpenAI's all-rounder: chat, images, voice, web search and agents in one app.
