ENفا
Get API key
MODELSArena’s interaction judge tests what agents actually finish▼ 22%DeepSeek V4.1 Flash API price#1Claude Fable 5.2 · 30.48% win shareFUNDINGAnthropic targets $64.4B Nasdaq IPO as revenue surgesTOOLSCursor tops the tools directory this weekMODELSGoogle cuts Gemini 4 Carbon prices before Gemini 2.5 sunsetPOLICYFive frontier AI labs face coordinated SF safety lawsuits
Vigas Arena · updated just now

The best Aligned models,
ranked by real use.

Evaluates nuanced safety, refusal calibration, honest uncertainty handling, and helpfulness under ambiguous instructions.

378,706,203sessions rankedHow we rank →

Rising this week · Aligned

GPT-6 Astraup 2.70 pts · #2 in AlignedGPT-6 Lunaup 2.37 pts · #3 in AlignedClaude Opus 5.5up 1.75 pts · #4 in Aligned

The podium · Aligned

Win share · this week
  1. Rank 1: GPT-6.1 SolOpenAI · 1488 score21.40%
  2. Rank 2: GPT-6 AstraOpenAI · 1485 score21.10%
  3. Rank 3: GPT-6 LunaOpenAI · 1482 score20.90%

Full ranking

Aligned leaderboard: win share, score, cost per task and votes
#ModelVotes7 days
1GPT-6.1 Sol (Max)OpenAI21.40%1488 ±2$1.03148,000
2GPT-6 Astra (Max)OpenAI21.10%1485 ±1$4.10144,733
3GPT-6 Luna (Low)OpenAI20.90%1482 ±1$0.15136,466
4Claude Opus 5.5 (Max)Anthropic18.40%1461 ±2$5.71139,199
5Grok 4.7xAI17.90%1455 ±1$1.85118,932
6Gemini 4 Carbon (High)Google17.20%1449 ±2$1.42129,665
7Claude Sonnet 5.5 (Max)Anthropic16.10%1441 ±2$1.95115,398
#121.40%
Score 1488 ±2$1.03/task148,000 votes
#221.10%
Score 1485 ±1$4.10/task144,733 votes
#320.90%
Score 1482 ±1$0.15/task136,466 votes
#418.40%
Score 1461 ±2$5.71/task139,199 votes
#517.90%
Score 1455 ±1$1.85/task118,932 votes
Score 1449 ±2$1.42/task129,665 votes
#716.10%
Score 1441 ±2$1.95/task115,398 votes

Pareto frontier

score vs $/task

Green models give the best score for their price. Tap any dot to inspect value vs. performance.

GPT-6.1 SolPareto optimal
Score 1488 · 21.40% win · $1.03/task
API pricing

Live sessions

LIVE

No sessions in the last 24 hours yet. Send a request through the Vigas API.

Connect via API — start in 30s
§ Methodology

How the rankings are made.

Real sessions, not benchmarks

Win share counts how often a model finishes real tasks that users accept — across real API workloads, coding agents, and blind Arena tests.

Blind pairwise votes

In Arena, people compare two hidden models and vote. Scores use a Bradley–Terry model with 95% confidence intervals.

Cost is part of the score

Every ranking shows dollars per completed task, so you can see who is best for the money — not just best.

§ Aligned · Questions & explore

About the aligned board.

Voice & conversational toolsPolicy newsAPI pricesAPI quickstart

What is the #1 AI model on the Aligned leaderboard right now?

GPT-6.1 Sol (OpenAI) leads aligned with a 21.40% real-work win share and a Bradley-Terry score of 1488, ahead of #2 GPT-6 Astra (21.10%).

Which top-10 model offers the lowest cost per task for aligned?

GPT-6 Luna balances #3 performance (score 1482) at $0.15 per completed task on Vigas.

How are Vigas leaderboard scores and win shares calculated?

Vigas combines real session outcomes across API workloads and agent harnesses with blind pairwise Arena votes using a Bradley-Terry statistical model with 95% confidence intervals.

Can I call the top aligned models through one API?

Yes. Point any OpenAI-compatible client at https://api.vigas.ai/v1 with your Vigas API key or set model to vigas/auto to route across providers automatically.