Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
Agentic benchmark
Artificial Analysis Harvey LAB-AA v1.0. Every model the catalog carries a published AA Harvey LAB v1.0 value for, ranked by that value.
Archived v1.0 legal-agent results from Artificial Analysis, kept separate from v1.1. 18 models in the ModelChoice catalog have a published value.
Each point is a model with a published AA Harvey LAB v1.0 value and complete token prices, on the 75% input / 25% output blend. Frontier models are spaced evenly along the cost axis.
18 models with a published AA Harvey LAB v1.0 value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.
| Rank | Model | Provider | Published v1.0 score |
|---|---|---|---|
| 1 | Moonshot AI | 94.6 | |
| 2 | Anthropic | 93.6 | |
| 3 | Anthropic | 93.5 |
What the benchmark measures and how its value is published, from the source’s own record.
AA Harvey LAB v1.0 (Artificial Analysis Harvey LAB-AA v1.0) is an agentic benchmark published by Artificial Analysis Harvey LAB-AA Benchmark Leaderboard. Archived v1.0 legal-agent results from Artificial Analysis, kept separate from v1.1. Its evaluation set is described by the source as legal agent tasks. The source rates its difficulty as professional legal work.
Each model's result is published by Artificial Analysis Harvey LAB-AA Benchmark Leaderboard as published v1.0 score. ModelChoice prints that value to 1 decimal place, exactly as published, and never converts or rescales it. This page ranks the 18 models that carry a published value, highest value first; a model the source never scored is left out rather than ranked at zero. The source does not state whether a higher value is the better result, so neither does this page.
| 4 | StepFun | 93.4 |
| 5 | Anthropic | 93.1 |
| 6 | Anthropic | 93.0 |
| 7 | xAI | 92.4 |
| 8 | Anthropic | 91.2 |
| 9 | 90.7 |
| 10 | Anthropic | 89.9 |
| 11 | MiniMax | 88.4 |
| 12 | OpenAI | 87.9 |
| 13 | OpenAI | 87.2 |
| 14 | OpenAI | 85.2 |
| 15 | NVIDIA | 81.7 |
| 16 | Mistral | 69.1 |
| 17 | xAI | 19.6 |
| 18 | OpenAI | 13.9 |
Evidence key: ObservedLast good
Rows are ordered by the value Artificial Analysis Harvey LAB-AA Benchmark Leaderboard published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.