Model Council asks three AI models and gives you one checked answer

Try it

Pick Model Council in the mode menu under the message box, then send a plan you want someone to argue with.

Every AI model has blind spots, and you cannot see them from inside one answer. A confident wrong answer reads exactly like a confident right one. The usual fix is to ask a second model, which means copying your question between tabs and comparing the replies by hand.

Model Council does that comparison for you. One message goes to three models at once, a separate judge lines their answers up, and you get one final answer along with the three it came from.

Three models from three different labs

The models in a council come from different companies, because models trained by different teams on different data tend to go wrong in different places. When all three agree, you can lean on the answer. When one disagrees, that disagreement is usually the part worth checking.

Today the council is:

• Standard: Claude Haiku 5.5, GPT-6 Luna and GLM 5.3 Flash
• Premium: Claude Haiku 5.5, GPT-6.1 Sol and GLM 5.3 Flash

We change the lineup as better models ship, and this page always lists the current one.

What comes back

Each answer gets its own tab, Answer 1, Answer 2 and Answer 3, and the tabs fill in as each model finishes. Under them, a synthesis panel shows:

• where the three answers agree
• where they disagree
• what each one found that the others missed

Then you get one final answer that keeps the strongest information from all three.

A judge that wrote none of the answers

The comparison is done by GPT-6 Luna Pro, which sits outside the council, so it never grades its own work. When you ask for a file or an image, such as a slide deck, a PDF or a logo, each council model makes its own version. The judge opens all three, picks the best one, marks it Best and says why.

The final answer is written by Claude Haiku 5.5 in Standard and GPT-6.1 Sol in Premium, both with high reasoning, working from the three answers and the judge's comparison.

When a council is worth it

Model Council pays for three answers plus the combined result, so it uses a lot more credits than a normal message. Save it for questions where a second opinion matters:

• Settling an argument: who is right, and what each side is missing.
• Checking a suspicious message: whether it is a scam, and the red flags each model spots.
• Stress-testing a plan: what breaks first, from three angles.
• Decisions with money attached: for a single clear recommendation instead, the AI decision maker is faster and cheaper.

You can also start one from the Model Council card on the create screen. It needs a free account, and it is not offered in group chats, which send a single answer.

The same idea, without the setup

Developers have been building "LLM councils" for a while: open source projects that send one question to several models and have a chairman model merge the replies. They work, but you need an account and an API key with each provider, and you read the results in a terminal. askROI runs the same pattern in an ordinary chat, with the lineup picked and kept current for you.

One model gives you an answer. Three models from different labs, compared by a judge that wrote none of them, also tell you how far to trust it.