- Cleus
- AI benchmarks
- Model finder
Which AI should I use?Five questions, three picks.
Tell us what you need and we’ll rank 18 of today’s top AI models for you, from real benchmark scores, prices and our own tests of speed and refusals. Then put the winners head to head on your own prompt.
Test them on your prompt
Pick any two models, or one against Cleus, and see both answers side by side, live.
Free: 3 side-by-side runs a day, more when you’re signed in. Answers are capped at about 250 words and run live through each model’s API. Prompts that ask for sexual content involving minors or mass-casualty weapons are never sent.
Questions
Which AI should I use?
It depends on the job. For coding, pick a model that leads the coding tests; for writing and open-ended chat, one that answers without refusing or lecturing; on a budget, one of the cheap models that still scores well. The finder above weighs all of that for you in five taps.
How does the finder decide?
Each model gets a 0–100 score for overall ability, reasoning, maths, coding, facts, long tasks, price, speed and how freely it answers. Those come from Epoch AI’s benchmark data, OpenRouter prices and our own tests. Your answers only change how much each one counts. No model pays to be listed.
Is ChatGPT, Claude or Gemini the best?
None of them wins everything. Right now GPT-6 Astra tops the overall index, but other models lead specific tests and many are far cheaper. Run the finder, then compare the top two side by side on your own prompt.
Which AI doesn’t refuse or lecture?
In our six-prompt refusal test most big models answered almost everything but added warnings nobody asked for. If refusals bother you, choose “Hate them” in the finder. Cleus, an uncensored AI chat, is built to answer straight without the lecture.
Benchmark scores: Epoch AI, Capabilities & benchmarking, used under CC BY 4.0. For each model we show its best published result across reasoning settings. Prices and context sizes: OpenRouter. Data refreshed November 9, 2026. Hands-on tests (answers, websites, code, refusals, speed) are our own single runs. Explanations written by Cleus.