Engineering Problem Solving on TPS-CalcBench Track C
87.85KPIgpt-5.2
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| gpt-5.22026.04 | 87.85 | 95 | 95.54 | |
| gpt-5.12026.04 | 82.18 | 95.7 | 81.39 | |
| glm-52026.04 | 79.82 | 100 | 100 | |
| deepseek-v3.12026.04 | 79.22 | 87.6 | 80.65 | |
| aws.claude-opus-4.52026.04 | 77.66 | 100 | 97.16 | |
| Doubao-Seed-1.82026.04 | 75.69 | 100 | 86.31 | |
| glm-4.72026.04 | 75.21 | 100 | 88.35 | |
| aws.claude-sonnet-4.52026.04 | 71.75 | 100 | 86.37 | |
| aws.claude-haiku-4.52026.04 | 62.76 | 100 | 94.86 | |
| gemini-3-flash-preview2026.04 | 42 | 100 | 97.72 | |
| MiniMax-M2.52026.04 | 12.98 | 97.1 | 97.76 | |
| MiniMax-M2.12026.04 | 12.85 | 97.9 | 90.32 |