Harvey's Legal Agent Benchmark, public 1,251-task set, all-pass (Harvey Legal Agent (public 1,251))
We show this table for reference; we do not rank on it.
Legal-agent tasks from the public 1,251-task Harvey dataset, scored by the share of tasks that pass every rubric criterion.
Benchmark score on Harvey Legal Agent (public 1,251) — October 29, 2026
We compile the Harvey Legal Agent (public 1,251) rows from provider self-reports. Kimi K3.1 Pro leads the table at 23.7%. We do not use these results to rank models overall.
1 modelAgenticCurrentDisplay onlyUpdated October 29, 2026
Benchmark score table (1 model)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | Kimi K3.1 ProMoonshot AI | 23.7% | Not reported | Pending |
About Harvey Legal Agent (public 1,251)
Year
2026
Tasks
1,251 public legal tasks
Format
All-pass rate
Difficulty
Professional legal work
Moonshot runs the public set with Artificial Analysis's Stirrup harness. It is a different task set from Harvey LAB-AA's 120 private tasks, so BenchLM keeps it on its own key.
Freshness and provenance
Version
Harvey Legal Agent (public 1,251) 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does Harvey Legal Agent (public 1,251) measure?
Legal-agent tasks from the public 1,251-task Harvey dataset, scored by the share of tasks that pass every rubric criterion.
Which model scores highest on Harvey Legal Agent (public 1,251)?
Kimi K3.1 Pro by Moonshot AI currently leads with a score of 23.7% on Harvey Legal Agent (public 1,251).
How many models are evaluated on Harvey Legal Agent (public 1,251)?
1 AI models have published results on Harvey Legal Agent (public 1,251) in the BenchLM catalog.
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 6,100+ subscribers.
One email each week. Unsubscribe anytime.