Compare / GPT-6 Sol vs GLM 5.3 Flash
GPT-6 Sol vs GLM 5.3 Flash
18 results from 4 sources that measured both models. Each row is in the source's own units; there is no overall winner.
| Benchmark | Metric | GPT-6 Sol | GLM 5.3 Flash |
|---|---|---|---|
| APEX-Agentsreported | Tasks passed% of tasks | 54.3% | 52.8% |
| Arena (formerly LMArena)reported | Confirmed task successIPS effect estimate | 0.11 | 0.07 |
| Arena (formerly LMArena)reported | Praise over complaintIPS effect estimate | 0.11 | -0.00 |
| Arena (formerly LMArena)reported | SteerabilityIPS effect estimate | 0.15 | -0.01 |
| Arena (formerly LMArena)reported | Tool groundingIPS effect estimate | 0.00 | 0.00 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,424 | 1,467 |
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,465 | 1,463 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,451 | 1,433 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,507 | 1,510 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,447 | 1,474 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,457 | 1,474 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,440 | 1,450 |
| Artificial Analysisreported | Artificial Analysis Intelligence Indexindex score | 47.6 | 41.8 |
| Artificial Analysisreported | Humanity's Last Exam% of questions | 47.9% | 39.9% |
| Artificial Analysisreported | Long-context reasoning (AA-LCR)% of questions | 83.7% | 80.0% |
| Artificial Analysisreported | SciCode% of problems | 57.6% | 51.6% |
| Artificial Analysisreported | Terminal-Bench 4.0% of tasks | 43.9% | 32.8% |
| GDP.pdfreported | Professional document tasks% of rubric | 26.4% | 14.0% |
Bold green marks the better value on that metric. Where a source reports ranges that overlap, the difference may not be meaningful; see the benchmark page for ranges.