Conversation modesEach mode changes how the models interact — same AIs, different behavior. Choose the mode that fits your question.

Alethe

Free

Alethe AI's default mode. Models read your question, assess what it requires, and respond naturally — a direct conversation if the question is simple, a structured debate if it's complex or controversial. They share a common transcript: each one sees what the others said before speaking. The result is a collectively built answer, more complete and less biased than any single model alone.

Example question"What's the best strategy to launch a B2B SaaS in 2025?"

Friendly

Free

Models switch to a warm, relaxed, and educational register. No technical jargon, no academic phrasing, no cold responses. Every explanation is clear, vivid, and accessible to anyone. Perfect for learning something new, understanding a complex concept without feeling lost, or just having a light conversation. The inter-model debate is still active under the hood — you get multiple perspectives, just in human language.

Example question"What actually is machine learning? Explain it like I'm 15."

Business

Pro

Models take the lens of a senior consultant: every answer is results-oriented, quantified when possible, and focused on real trade-offs. They analyze risks, size opportunities, compare options using ROI and impact criteria. No hollow theory, no generalities — only actionable recommendations tailored to your context. Ideal for strategic decisions, investment choices, product pivots, or growth planning.

Example question"Should we raise funding now or stay bootstrapped?"

Wellness

Pro

Models explore your everyday health questions — nutrition, sleep, exercise, stress management, and healthy habits — with rigor and nuance. They distinguish what is established (studies, consensus) from what remains debated or individual-dependent. Every response includes a reminder that AI models are not a substitute for a healthcare professional — this mode helps you understand, build better habits, and ask the right questions at your next appointment.

Example question"How can I naturally improve my sleep quality?"

Socratic

Ultra

Inspired by Socratic maieutics: models never give the answer directly. Instead, they ask precise questions that force you to articulate your thinking, identify your contradictions, and build your own reasoning. Each AI brings a different angle — one challenges your assumptions, another explores the consequences, a third uncovers what you haven't yet put into words. The truth you find yourself is always more solid than the one you're handed.

Example question"I want to decide whether I should change careers."

Debate

Ultra

Models take opposing positions and defend them with argumentative rigor — even if they don't personally hold those views. No soft consensus, no forced convergence: each AI maintains its stance against the others' counter-arguments, cites facts, anticipates objections, and dismantles opposing reasoning. Designed to explore every facet of a controversial question, understand the strengths of each side, and forge your own opinion on solid ground.

Example question"Is full remote work good or bad for companies?"

Custom

Apex

You define the models' behavior entirely through a system prompt you write yourself. Give them a precise role (legal expert, life coach, financial analyst…), specific constraints, a particular personality, a domain of expertise, or an expected response format. The models follow your instructions while continuing to debate each other within that framework. Before launch, an AI evaluates your prompt out of 100 across 4 criteria: multi-AI relevance, debate potential, flow, and language clarity.

Example question"You are a French tax law expert specializing in startups. Analyze each situation with legal rigor and cite specific articles of law."

An AI scores your prompt (0–100) across 4 criteria: multi-AI suitability, debate potential, flow, and language clarity — before you launch.

How to write an effective prompt

Mode availability by plan

Free

Alethe, Friendly

Pro

+ Business, Wellness

Ultra

+ Socratic, Debate

Apex

+ Custom (your own prompt)

See plans

35 models across 4 tiersYou choose which models participate in each conversation. Access depends on your plan — Free unlocks Tier 1, Apex unlocks all 35.

Tier 1 — Free

Gemini 2.5 Flash

DeepSeek V3.2

Gemini 3.1 Flash Lite

Gemini 2.0 Flash

Gemini 3 Flash

Llama 4 Maverick

Llama 4 Scout

Llama 3.3 70B

Qwen 3.6 Plus

Qwen 2.5 72B

Mistral Small

Grok 3 Mini

Tier 2 — Pro

GPT-5 Mini

GPT-4o Mini

GPT-4.1

DeepSeek R1

Claude Haiku 4.5

Mistral Medium

Mistral Large

MiMo V2 Pro

Tier 3 — Ultra

Claude Sonnet 4.6

Claude Sonnet 4.5

Gemini 3.1 Pro

GPT-5.4

GPT-4o

Grok 3

Grok 4

o4 Mini

Tier 4 — Apex

GPT-5

GPT-5.4 Pro

o3 Pro

Claude Opus 4.6

How it works