> ## Documentation Index
> Fetch the complete documentation index at: https://docs.dualmindlab.tech/llms.txt
> Use this file to discover all available pages before exploring further.
# Welcome to DualMind Arena
> The blind AI model comparison platform. Compare leading AI models side-by-side, vote on the best response, and discover model quality through crowd-sourced benchmarks.
## What is DualMind Arena? v2.0
DualMind Arena is a **blind AI model comparison platform** — submit a prompt, receive responses from two competing models simultaneously, vote on the better one, and contribute to a community-driven ELO-based leaderboard.
The key insight: **knowing a model's name changes how you judge it.** DualMind hides model identities until after you vote — so quality determines the ranking, not brand recognition.
Submit one prompt to two AI models simultaneously. Vote blind. See the truth.
Every vote shifts real ELO ratings. The leaderboard reflects collective human preference, not marketing budgets.
Organize comparisons into persistent threads. Share publicly, keep private, or distribute via link.
Time to First Token and Tokens/Second tracked for every response. Speed is measured separately — never folded into quality.
***
## How it works
```mermaid theme={null}
sequenceDiagram
participant You
participant Arena as DualMind Arena
participant A as Model A (hidden)
participant B as Model B (hidden)
You->>Arena: Submit a prompt
Arena->>A: Forward prompt
Arena->>B: Forward prompt (parallel)
A-->>Arena: Stream response
B-->>Arena: Stream response
Arena-->>You: Display both responses (blind)
You->>Arena: Vote — A, B, or Tie
Arena-->>You: Reveal model identities
Arena-->>You: Update ELO rankings live
```
***
## Platform modes
Two randomly-selected models. One prompt. Zero brand bias.
Responses appear side-by-side under anonymous labels — **Model A** and **Model B**. Vote for the better response, then see which models you were actually comparing.
This is the primary mode for leaderboard contributions. Every vote carries statistical weight in the ELO system.
Direct conversation with a specific model of your choice.
Useful when you want to evaluate a known model in depth, test specific capabilities, or have a sustained conversation without the comparison framing.
Organize your comparisons into named threads with full history.
* **Private** — visible only to you
* **Unlisted** — accessible via link, not listed publicly
* **Public** — discoverable by anyone
Perfect for research documentation, sharing interesting comparisons, or building a personal evaluation archive.
***
## What makes this different
In every study on AI evaluation, knowing a model's identity introduces measurable bias. Users consistently rate GPT-4 responses higher when they know it's GPT-4 — even when the content is identical to a competitor's output.
Blind testing removes this entirely at the architecture level. Model names are never sent to the client until after a vote is submitted. This is enforced as a data contract, not a UI convention.
Win rate is a static snapshot. A model with a 60% win rate against weak opponents tells you nothing about how it performs against the best.
ELO is dynamic. It adjusts based on the strength of who you beat. A model that defeats high-ranked competitors gains more points than one that beats weak ones. This produces a leaderboard that reflects true relative quality, not raw vote counts.
A faster model response creates an immediate impression of fluency, even before the user reads a word. This is well-documented in UX research.
Latency is infrastructure — it reflects hosting, not intelligence. The Arena displays Time to First Token (TTFT) and Tokens per Second (TPS) as separate, transparent metrics — never folded into a quality score.
***
## Start here
Run your first comparison in under two minutes.
Deep dive into blind comparison, ELO scoring, and streaming architecture.
The reasoning behind how we design fair, meaningful AI comparisons.
What we're building next and why.
Ready? Head to the [Quickstart guide](/quickstart) and run your first comparison.