Automated Inequality Proving on PolyIneqBench n=7
26.7Pass RateNSPI
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| NSPI2026.05 | 26.7 | 29.6 | |
| Gemini-3-Pro2026.05 | 15 | 140.7 | |
| LIPS2026.05 | 8.3 | 150.2 | |
| Maple2026.05 | 6.7 | 821.5 | |
| DeepSeek-V3.22026.05 | 5 | 393.8 | |
| GPT-5.22026.05 | 3.3 | 75.3 | |
| Z32026.05 | 1.7 | 65 | |
| Kimina-Prover2026.05 | 1.7 | 120.6 | |
| DS-Prover-v22026.05 | 0 | — | |
| Goedel-Prover-v22026.05 | 0 | — |