Benchmark Registry

HLAB

Harvey's Legal Agent Benchmark

Updated

Benchmark
HLAB
Evaluated by
Vals AI
Release date
October 6, 2026
Version
HLAB October 2026 evaluation
Metric
HLAB Task Pass Rate
Cite

Fill in the access date before citing.

Plain text 

Benchmark Registry. Harvey's Legal Agent Benchmark October 2026 evaluation benchmark results. Benchmark identifier: harveys-legal-agent-benchmark/october-2026. https://benchmarkregistry.org/benchmarks/harveys-legal-agent-benchmark/october-2026. Accessed [YYYY-MM-DD]. Data license: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Attribution: Benchmark Registry, a project of Densa Labs — https://benchmarkregistry.org/ (CC BY 4.0). Include the cited record URL and identify any changes.

BibTeX 

@misc{br_harveys_legal_agent_benchmark_october_2026,
  title = {Harvey's Legal Agent Benchmark October 2026 evaluation benchmark results},
  publisher = {Benchmark Registry},
  url = {https://benchmarkregistry.org/benchmarks/harveys-legal-agent-benchmark/october-2026},
  urldate = {YYYY-MM-DD},
  note = {Benchmark identifier: harveys-legal-agent-benchmark/october-2026; data license: CC BY 4.0; access date to be filled by the reader}
}
Report an issue

Results

50 results

ProviderModelScoreSourceRegistry No.
OpenAIGPT-6 Astra (max)5.4%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord10005
OpenAIGPT-5.4 (xhigh)0.0%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord10007
OpenAIGPT-5.5 (xhigh)3.8%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord10008
OpenAIGPT-5.6 Luna (max)1.3%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord10009
OpenAIGPT-5.6 Sol (max)2.5%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord10010
OpenAIGPT-5.6 Terra (max)0.8%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord10011
OpenAIGPT-6 Luna (max)2.9%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord10013
OpenAIGPT-6 Sol (max)1.7%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord10014
OpenAIGPT-6.1 Sol (max)5.4%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord10015
OpenAIGPT-5.4 mini (xhigh)0.0%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord10021
OpenAIGPT-5.4 nano (high)0.0%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord10022
DeepSeek AIDeepSeek-V4-Pro (max)3.8%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord110002
Moonshot AIKimi K2.5 (thinking)0.0%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord120001
Moonshot AIKimi K2.61.7%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord120002
Moonshot AIKimi K3 (max)12.9%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord120003
Alibaba (Tongyi)Qwen3.6-Plus1.3%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord130004
Alibaba (Tongyi)Qwen3.7-Max1.7%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord130006
Alibaba (Tongyi)Qwen3.7-Plus0.0%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord130007
Alibaba (Tongyi)Qwen3.8-Max10.4%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord130008
MiniMaxMiniMax M2.70.0%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord140002
MiniMaxMiniMax M34.2%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord140003
Z.aiGLM-5.10.0%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord150003
Z.aiGLM-5.2 (max)7.1%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord150004
Z.aiGLM-5.3 (max)8.3%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord150005
Z.aiGLM-5.3-Flash (max)6.7%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord150006
Thinking Machines LabInkling (0.99)2.1%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord160001
Thinking Machines LabInkling-Small (0.99)1.7%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord160002
AnthropicClaude Fable 5.16.7%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord20004
AnthropicClaude Haiku 4.5 (thinking)0.8%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord20006
AnthropicClaude Sonnet 4.65.0%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord20009
AnthropicClaude Opus 4.76.7%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord20010
AnthropicClaude Opus 4.89.6%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord20011
AnthropicClaude Sonnet 55.0%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord20013
AnthropicClaude Opus 56.7%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord20014
AnthropicClaude Opus 5.53.8%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord20015
AnthropicClaude Sonnet 5.52.9%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord20016
Google (DeepMind)Gemini 3.8 Flash (high)10.0%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord30004
Google (DeepMind)Gemini 3.1 Pro (high)0.0%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord30005
Google (DeepMind)Gemini 3.5 Flash (high)2.5%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord30006
Google (DeepMind)Gemini 3.6 Flash (high)3.3%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord30007
Google (DeepMind)Gemini 3 Flash (high)0.0%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord30009
Google (DeepMind)Gemini 3.7 Flash (high)8.8%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord30010
SpaceXAIGrok 4.5 (high)12.9%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord40001
SpaceXAIGrok 4.6 (high)15.8%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord40002
SpaceXAIGrok 4.7 (xhigh)12.5%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord40003
NVIDIANVIDIA Nemotron 3 Ultra 550B-A55B0.4%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord60003
Meta AI (originally Facebook AI Research)Muse Spark 1.1 (xhigh)20.0%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord80002
Meta AI (originally Facebook AI Research)Muse Spark 1.2 (xhigh)25.4%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord80003
Meta AI (originally Facebook AI Research)Muse Spark 1.3 (xhigh)22.9%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord80005
Mistral AIMistral Medium 3.5 (high)0.4%vals.aiArchived copyHLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord90005