PolicyBench
GPT-5.5
OpenAI · AI alone, no tools
United States rank
#3 of 23
US exact
83.5%
United States
United States benchmark
Exact match
83.5%
Bounded score
93.3%
Parse rate
100.0%
Eligibility flags
945/984
Score by programhardest first
| Program | Exact | Within 1% | n |
|---|---|---|---|
| State tax before refundable credits | 60.0% | 67.0% | 100 |
| Federal tax before refundable credits | 61.0% | 65.0% | 100 |
| SNAP | 76.0% | 80.0% | 100 |
| State refundable credits | 83.0% | 84.0% | 100 |
| Federal refundable credits | 88.0% | 90.0% | 100 |
| Payroll tax | 88.0% | 90.0% | 100 |
| Person-level Medicare eligibility | 89.8% | 89.8% | 177 |
| Person-level CHIP eligibility | 94.9% | 94.9% | 177 |
| Person-level Medicaid eligibility | 94.9% | 94.9% | 177 |
| SSI | 95.0% | 96.0% | 100 |
| Free school meals eligibility | 98.0% | 98.0% | 100 |
| Reduced-price school meals eligibility | 99.0% | 99.0% | 100 |
| TANF | 99.0% | 99.0% | 100 |
| Local income tax | 100.0% | 100.0% | 100 |
| Person-level Early Head Start eligibility | 100.0% | 100.0% | 38 |
| Person-level Head Start eligibility | 100.0% | 100.0% | 38 |
| Person-level WIC eligibility | 100.0% | 100.0% | 177 |
| Self-employment tax | 100.0% | 100.0% | 100 |
Hardest casesworst misses on positive references
- State tax before refundable credits2842% offReference $7 · predicted $200Inspect household #023
- SNAP675% offReference $461 · predicted $3,576Inspect household #023
- State tax before refundable credits287% offReference $514 · predicted $1,990Inspect household #039
- Federal tax before refundable credits282% offReference $1,346 · predicted $5,145Inspect household #039
- Federal tax before refundable credits132% offReference $5,200 · predicted $12,040Inspect household #123
- Federal refundable credits100% offReference $2,013 · predicted $0Inspect household #026
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.