TipWhich AI is smartest right now? New independent ranking: Claude Opus 5.5 ahead of GPT-6 Astra, and the cheapest top model costs 13 times less
Analysis firm Artificial Analysis has rebuilt its big AI test. In the new Intelligence Index v5, Claude Opus 5.5 leads with 58 points ahead of GPT-6 Astra at 57. Gemini 4 Argon scores 50. Among the top ten, GPT-6.1 Terra is cheapest per task, more than 13 times cheaper than the most expensive. Here is how to use the list when choosing an AI.
Artificial Analysis is an independent company that tests AI models and compares price and performance. On 27 October 2026 it released version 5 of its Intelligence Index, a combined score built on nine different tests. They measure four equally weighted areas: knowledge work (such as producing reports and automating workflows), coding, science and general knowledge. For the first time, 60 percent of the test uses private questions, so AI companies cannot have trained on the answers in advance.
The top looks like this: Anthropic’s Claude Opus 5.5 leads with 58 points, followed by OpenAI’s GPT-6 Astra at 57 and GPT-6.1 Sol at 55. Anthropic and OpenAI hold the five best models. Google’s Gemini 4 Argon comes next at 50. No model wins all nine tests: Claude Opus 5.5 wins three, and five other models share the rest. The tests run at the highest reasoning level, which is not always how you meet the model in a free app.
Open models, which anyone can download and run themselves, have fallen behind. Kimi K3.1 Pro scores 46 but is not yet downloadable. Among models you can actually download, DeepSeek V4.1 Flash and GLM-5.3 lead with 32 points, below Anthropic’s low-cost Claude Haiku 5.5 at 36.
The price gap is big. Among the ten highest-scoring models, GPT-6.1 Terra costs least, 0.96 dollars per test task, while Claude Fable 5.1 costs 13.04 dollars. That is more than 13 times as much. If you pay per use, for example a business with its own AI automation, a model a few points down the list may be the smartest choice.
Good to know: the test is in English and mostly text-based. Artificial Analysis measures images, speech and other languages separately, so the list does not say how good a model is in Swedish, Norwegian, Danish or Finnish. Nor does it say what is included in your plan. Test two or three models on a real task from your own work before switching, and see how we rank AI tools for what they cost.
Update 29 October: the day after the list came out, Google’s new Gemini 4 Carbon took first place with 59 points, according to Artificial Analysis. Carbon cannot be used yet, except by a small group of security experts. Read more in our article on Gemini 4 Carbon.
How to get started
- Decide what you mostly use AI for: writing, calculating, coding or doing tasks for you.
- Pick two or three models from the top of the list that are in plans you can afford.
- Give them the same real task, ideally in your own language, and compare the results.
- If you pay per token: work out what the task costs before choosing the top-ranked one.
More on TipAI
Sources
- Artificial Analysis: Announcing Intelligence Index v5 (27 okt 2026) (retrieved 2026-10-28)
- Artificial Analysis: Intelligence Benchmarking, metod v5 (retrieved 2026-10-28)
- OfficeChai: Artificial Analysis Launches Intelligence Index v5 (28 okt 2026) (retrieved 2026-10-28)
- Artificial Analysis: Gemini 4 Carbon is the new #1 on the Intelligence Index v5 (28 okt 2026) (retrieved 2026-10-29)
Free cheat sheet: Get started with AI
A printable page with the basic formula, five things to try and how to use AI safely. Enter your email and we will send it straight away – free, unsubscribe any time.
Get the cheat sheet free →TipAI explains AI in plain language for beginners, families and small businesses. Every article has sources and a "How to get started" section.