Skip to main content
PolicyBench

Qwen 3.7 Max

Alibaba · AI alone, no tools
Current board, snapshot 2026-11-11
United States rank
#43 of 52
US exact
76.9%
United States

United States benchmark

Exact match
76.9%
Bounded score
86.5%
Parse rate
100.0%
Eligibility flags
923/975

Exact match by programall 18 output groups, best first

Bar length is the unweighted exact-match rate on that program's outputs; the right column is the program's share of the headline weight, which follows household dollars, so the bars do not average to the headline.

  1. Local income tax100.0%0.3%
  2. Reduced-price school meals eligibility99.0%0.1%
  3. TANF99.0%0.3%
  4. Person-level WIC eligibility98.9%0.3%
  5. Person-level Medicare eligibility98.3%11%
  6. Person-level Early Head Start eligibility97.4%3.1%
  7. SSI96.9%2.0%
  8. Free school meals eligibility96.0%0.7%
  9. Self-employment tax96.0%2.1%
  10. Person-level CHIP eligibility90.4%0.2%
  11. Person-level Head Start eligibility89.5%1.2%
  12. Person-level Medicaid eligibility88.4%30%
  13. SNAP80.6%4.1%
  14. State refundable credits80.6%0.6%
  15. Payroll tax76.0%16%
  16. Federal refundable credits76.0%3.7%
  17. State tax before refundable credits49.5%6.0%
  18. Federal tax before refundable credits44.0%19%
Table view (exact, within 1%, outputs), hardest first
ProgramExactWithin 1%n
Federal tax before refundable credits44.0%45.2%84
State tax before refundable credits49.5%49.5%93
Federal refundable credits76.0%77.0%100
Payroll tax76.0%83.3%96
State refundable credits80.6%80.6%98
SNAP80.6%80.6%93
Person-level Medicaid eligibility88.4%88.4%173
Person-level Head Start eligibility89.5%89.5%38
Person-level CHIP eligibility90.4%90.4%177
Free school meals eligibility96.0%96.0%100
Self-employment tax96.0%96.0%100
SSI96.9%96.9%97
Person-level Early Head Start eligibility97.4%97.4%38
Person-level Medicare eligibility98.3%98.3%172
Person-level WIC eligibility98.9%98.9%177
Reduced-price school meals eligibility99.0%99.0%100
TANF99.0%99.0%100
Local income tax100.0%100.0%98

Hardest casesworst misses on positive references

Scores are from the frozen manuscript snapshot under the AI-alone condition: structured responses over the same household facts (whole-scenario or in output subsets, per the serving-configuration table), no tools, graded against PolicyEngine reference outputs. See the leaderboard and paper for methodology. The frozen US annotations cover 9,132 rows whose legacy threshold score is below 1. That universe contains 9,128 of 9,128 exact-match misses and 4 exact hits. Another 2,497 rows with bounded score below 100 were not selected and have no audit annotation.