# Artificial Analysis Harvey LAB-AA (AA Harvey LAB)

> An independently evaluated legal-agent benchmark from Artificial Analysis.

Canonical page: https://benchlm.ai/benchmarks/aaharveylab

- Category: [Agentic](/agentic)
- Last updated: September 27, 2026

## About AA Harvey LAB

- Year: 2026
- Tasks: Legal agent tasks
- Format: Task success rate
- Difficulty: Professional legal work
- Paper: [Artificial Analysis Harvey LAB-AA Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/harvey-lab-aa)

Stored as its own agentic row.

BenchAlign v5.7 gives AA Harvey LAB 1% of the Agentic reference weight, so it moves the Agentic leaderboard and the overall ranking. Reference weights are relative weights in the calibrated model, not fixed shares of a score.

## Leaderboard (12 models)

| Rank | Model | Creator | Score |
|------|-------|---------|-------|
| 1 | [Kimi K3](/models/kimi-k3) | Moonshot AI | 94.6% |
| 2 | [Claude Fable 5](/models/claude-fable) | Anthropic | 93.6% |
| 3 | [Claude Opus 5](/models/claude-opus-5) | Anthropic | 93.5% |
| 4 | [Claude Fable 5.1](/models/claude-fable-5-1) | Anthropic | 93.0% |
| 5 | [Claude Opus 5.5](/models/claude-opus-5-5) | Anthropic | 91.2% |
| 6 | [Gemini 3.7 Flash](/models/gemini-3-7-flash) | Google | 90.7% |
| 7 | [MiniMax M3](/models/minimax-m3) | MiniMax | 88.4% |
| 8 | [GPT-5.6 Luna](/models/gpt-5-6-luna) | OpenAI | 87.9% |
| 9 | [GPT-5.6 Sol](/models/gpt-5-6-sol) | OpenAI | 87.2% |
| 10 | [Nemotron 3 Ultra](/models/nemotron-3-ultra) | NVIDIA | 81.7% |
| 11 | [Mistral Medium 3.5 128B](/models/mistral-medium-3-5-128b) | Mistral | 69.1% |
| 12 | [Grok 4.7](/models/grok-4-7) | xAI | 19.6% |

## FAQ

### What does AA Harvey LAB measure?

An independently evaluated legal-agent benchmark from Artificial Analysis.

### Which model scores highest on AA Harvey LAB?

Kimi K3 by Moonshot AI currently leads with a score of 94.6% on AA Harvey LAB.

### How many models are evaluated on AA Harvey LAB?

12 AI models have been evaluated on AA Harvey LAB on BenchLM.

### Does AA Harvey LAB affect BenchLM's overall score?

Yes. BenchAlign v5.7 gives AA Harvey LAB 1% of the Agentic reference weight, so it moves the Agentic leaderboard and the overall ranking. Reference weights are relative weights in the calibrated model, not fixed shares of a score.

## Compare Top Models on AA Harvey LAB

- [Kimi K3 vs Claude Fable 5](/compare/claude-fable-vs-kimi-k3)
- [Claude Fable 5 vs Claude Opus 5](/compare/claude-fable-vs-claude-opus-5)
- [Claude Opus 5 vs Claude Fable 5.1](/compare/claude-fable-5-1-vs-claude-opus-5)
- [Claude Fable 5.1 vs Claude Opus 5.5](/compare/claude-fable-5-1-vs-claude-opus-5-5)