BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases12

Harvey LAB-AA leaderboard

As of 9 Nov 2026, the top score on Harvey LAB-AA is 13.1% by Muse Wonder Full (Meta), ahead of Muse Wonder at 10.0%, across 33 model configurations with a published result (Harvey, result dated 12 Oct 2026). Legal work drawn from practice, built with Harvey. Run by Artificial Analysis. Harvey LAB-AA is shown for reference only and does not count toward the BenchLeader Index.

Key facts (as of 9 Nov 2026)

  • Muse Wonder Full holds the top Harvey LAB-AA v1.1 score at 13.1% (Harvey, 12 Oct 2026).
  • Muse Wonder is second on Harvey LAB-AA v1.1 at 10.0% (Harvey, 12 Oct 2026).
  • Harvey LAB-AA v1.1: Legal work drawn from practice, built with Harvey. Run by Artificial Analysis (Harvey).
  • 33 model configurations have a published Harvey LAB-AA v1.1 result on BenchLeader (BenchLeader, 9 Nov 2026).
  • Harvey LAB-AA v1.1 does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).

Legal work drawn from practice, built with Harvey. Run by Artificial Analysis.

Published by
Harveydata via Artificial Analysis
Category
Agents & tools
Index weight
Reference only
Models
33
Data as of
9 Nov 2026

Source: Artificial Analysis (artificialanalysis.ai). Data taken from the public leaderboard.

Top 15
  1. 1Muse Wonder Full (max)13.1%
  2. 2Muse Wonder (max)10.0%
  3. 3Grok 4.7 (xhigh)9.4%
  4. 4Muse Spark 1.3 (max)8.9%
  5. 5GPT-6 Astra (max)8.6%
  6. 6Claude Fable 5.2 (max)7.2%
  7. 7GPT-6.1 Sol (max)6.9%
  8. 8Claude Fable 5.1 (max)6.4%
  9. 9Claude Fable 5.2 (medium)6.4%
  10. 10Kimi K3 (max)5.3%
  11. 11Claude Fable 5.2 (high)4.7%
  12. 12Claude Opus 5.5 (max)4.2%
  13. 13Claude Fable 5.2 (low)4.2%
  14. 14Claude Fable 5.2 (xhigh)3.9%
  15. 15Gemini 4 Argon (high)3.9%
33 of 33
#
1Muse Wonder Fullmax13.1%62.22026-10-12
2Muse Wondermax10.0%61.92026-10-12
3Grok 4.7xhigh9.4%58.12026-09-21
4Muse Spark 1.3max8.9%61.72026-09-02
5GPT-6 Astramax8.6%67.42026-09-03
6NewClaude Fable 5.2max7.2%69.42026-11-09
7GPT-6.1 Solmax6.9%67.02026-09-29
8Claude Fable 5.1max6.4%66.92026-09-01
9NewClaude Fable 5.2medium6.4%66.32026-11-09
10Kimi K3max5.3%62.02026-07-16
11NewClaude Fable 5.2high4.7%68.92026-11-09
12Claude Opus 5.5max4.2%68.92026-09-22
13NewClaude Fable 5.2low4.2%65.32026-11-09
14NewClaude Fable 5.2xhigh3.9%68.72026-11-09
15Gemini 4 Argonhigh3.9%65.72026-09-30
16GPT-6 Solmax3.6%62.92026-09-22
17GPT-6 Lunamax3.3%56.52026-09-22
18Mistral Large 43.1%52.22026-10-06
19Claude Sonnet 5.5max2.8%65.02026-09-28
20Claude Haiku 5.5xhigh2.8%56.82026-10-07
21DeepSeek V4.1 Flashmax1.7%58.92026-09-10
22Claude Haiku 5.5low1.7%47.82026-10-07
23Claude Haiku 5.5high1.1%55.52026-10-07
24GLM 5.3 Flash0.8%56.82026-08-26
25Claude Haiku 5.5max0.6%56.82026-10-07
26Claude Haiku 5.5medium0.6%50.22026-10-07
27GLM 5.3max0.3%58.92026-08-18
28Gemini 3.8 Flashhigh0.0%61.82026-09-02
29Gemini 3.1 Pro0.0%60.62026-02-19
30MiniMax M30.0%56.32026-06-01
31Inkling Small0.0%52.12026-07-30
32Gemini 3.5 Flash Lite0.0%51.52026-07-21
33Inklingxhigh0.0%50.12026-07-15

More agents & tools benchmarks · All LLM benchmarks

Related benchmarks

Top head-to-heads on Harvey LAB-AA v1.1

Cite as: BenchLeader, “Harvey LAB-AA v1.1 leaderboard”, data as of 9 Nov 2026, https://www.benchleader.com/benchmarks/aa_harvey_lab.

Harvey LAB-AA v1.1: questions

What does Harvey LAB-AA v1.1 measure?
Legal work drawn from practice, built with Harvey. Run by Artificial Analysis. Scores are reported in percent of tasks solved; higher is better.
What is the top score on Harvey LAB-AA?
As of 9 Nov 2026, the highest score on Harvey LAB-AA v1.1 is 13.1% by Muse Wonder Full (result dated 12 Oct 2026), ahead of Muse Wonder at 10.0%. Scores come from Harvey and are re-read daily.
Which AI model leads Harvey LAB-AA v1.1?
Muse Wonder Full leads Harvey LAB-AA v1.1 with 13.1% as of 9 Nov 2026 (result dated 12 Oct 2026), ahead of Muse Wonder at 10.0%.
How many models have Harvey LAB-AA v1.1 results?
33 model configurations have a Harvey LAB-AA v1.1 result on BenchLeader, all taken from Harvey via Artificial Analysis.
Who runs Harvey LAB-AA v1.1 and how often is it updated?
Harvey LAB-AA v1.1 is published by Harvey. BenchLeader re-reads the published results once a day and records the date each result was published.
Does Harvey LAB-AA v1.1 count toward the BenchLeader Index?
No. Harvey LAB-AA v1.1 is shown for reference but left out of the composite index.