Multilingual Reasoning on OMH-Polyglot
99.5AccuracyRaw
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| RawCloud Agent=gpt-3.5-turbo2026.06 | 99.5 | 94,338 | 53,713 | 40,625 | 99.04 | |
| Cross-Lingual Token ArbitrageCloud Agent=gpt-3.5-turbo, Compression Rate=r=0.62026.06 | 99.5 | 86,474 | 28,661 | 57,812 | 99.08 | |
| Cross-Lingual Token ArbitrageCloud Agent=gpt-4o, Compression Rate=r=0.62026.06 | 99.5 | 127,594 | 28,776 | 98,819 | 98.88 | |
| RawCloud Agent=gpt-4o2026.06 | 98.33 | 139,085 | 43,565 | 95,520 | 97.66 | |
| Cross-Lingual Token ArbitrageCloud Agent=gemini-2.5-flash-lite, Compression Rate=r=0.62026.06 | 98 | 94,725 | 29,398 | 65,327 | 97.54 | |
| LLMLingua-2Cloud Agent=gpt-4o, Compression Rate=r=0.62026.06 | 97.17 | 124,790 | 37,281 | 87,509 | 96.56 | |
| RawCloud Agent=gemini-2.5-flash-lite2026.06 | 95 | 116,653 | 44,918 | 71,735 | 94.43 | |
| LLMLingua-2Cloud Agent=gpt-3.5-turbo, Compression Rate=r=0.62026.06 | 77.33 | 85,026 | 42,314 | 42,712 | 76.91 | |
| LLMLingua-2Cloud Agent=gemini-2.5-flash-lite, Compression Rate=r=0.62026.06 | 35 | 243,193 | 37,932 | 205,261 | 33.85 |