Compare / GLM 4.6 vs GLM 5.1
GLM 4.6 vs GLM 5.1
20 results from 3 sources that measured both models. Each row is in the source's own units; there is no overall winner.
| Benchmark | Metric | GLM 4.6 | GLM 5.1 |
|---|---|---|---|
| Arena (formerly LMArena)reported | OverallArena rating | 1,447 | 1,456 |
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,412 | 1,465 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,400 | 1,448 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,439 | 1,495 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,412 | 1,460 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,424 | 1,465 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,405 | 1,453 |
| Artificial Analysisreported | Artificial Analysis Coding Indexindex score | 45.8 | 55.8 |
| Artificial Analysisreported | Artificial Analysis Intelligence Indexindex score | 18.5 | 26.1 |
| Artificial Analysisreported | GPQA Diamond% of questions | 78.0% | 86.8% |
| Artificial Analysisreported | Humanity's Last Exam% of questions | 14.5% | 30.1% |
| Artificial Analysisreported | IFBench% of instructions | 43.4% | 76.3% |
| Artificial Analysisreported | Long-context reasoning (AA-LCR)% of questions | 54.0% | 73.7% |
| Artificial Analysisreported | Terminal-Bench 2.1% of tasks | 49.4% | 61.8% |
| Artificial Analysisreported | Terminal-Bench Hard% of tasks | 28.8% | 43.2% |
| Artificial Analysisreported | Τ-bench banking% of tasks | 13.4% | 13.6% |
| Artificial Analysisreported | Τ²-bench telecom% of tasks | 76.9% | 97.7% |
| UGI Leaderboardreported | Requested-length error% off the requested word count | 13.0% | 9.0% |
| UGI Leaderboardreported | Style adherencescore from 0 to 1 | 0.38 | 0.38 |
| UGI Leaderboardreported | Writing scorescore out of 100 | 53.3 | 61.7 |
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.