Skip to main content
PolicyBench

Kimi K2.6

Moonshot AI · AI alone, no tools
United States rank
#22 of 23
US exact
64.6%
United States

United States benchmark

Exact match
64.6%
Bounded score
67.0%
Parse rate
78.5%
Eligibility flags
869/984

Score by programhardest first

ProgramExactWithin 1%n
Federal tax before refundable credits40.0%40.0%100
Federal refundable credits45.0%45.0%100
Payroll tax46.0%46.0%100
State tax before refundable credits46.0%46.0%100
State refundable credits48.0%48.0%100
Self-employment tax60.0%60.0%100
SNAP76.0%76.0%100
Person-level Medicaid eligibility82.5%82.5%177
Person-level CHIP eligibility83.6%83.6%177
SSI86.0%87.0%100
Local income tax87.0%87.0%100
TANF89.0%89.0%100
Reduced-price school meals eligibility90.0%90.0%100
Person-level Medicare eligibility90.4%90.4%177
Person-level WIC eligibility91.0%91.0%177
Free school meals eligibility91.0%91.0%100
Person-level Early Head Start eligibility94.7%94.7%38
Person-level Head Start eligibility97.4%97.4%38

Hardest casesworst misses on positive references

Scores are from the frozen manuscript snapshot under the AI-alone condition: one structured response per household, no tools, graded against PolicyEngine reference outputs. See the leaderboard and paper for methodology.