Compare / DeepSeek V3.2 vs GPT-5.6 Luna
DeepSeek V3.2 vs GPT-5.6 Luna
16 results from 3 sources that measured both models. Each row is in the source's own units; there is no overall winner.
| Benchmark | Metric | DeepSeek V3.2 | GPT-5.6 Luna |
|---|---|---|---|
| Arena (formerly LMArena)reported | OverallArena rating | 1,437 | 1,461 |
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,423 | 1,463 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,400 | 1,412 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,453 | 1,491 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,420 | 1,446 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,425 | 1,454 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,410 | 1,435 |
| Artificial Analysisreported | Artificial Analysis Coding Indexindex score | 44.2 | 71.4 |
| Artificial Analysisreported | Artificial Analysis Intelligence Indexindex score | 21.5 | 37.3 |
| Artificial Analysisreported | GPQA Diamond% of questions | 84.0% | 91.1% |
| Artificial Analysisreported | Humanity's Last Exam% of questions | 24.6% | 39.5% |
| Artificial Analysisreported | Long-context reasoning (AA-LCR)% of questions | 73.3% | 83.7% |
| Artificial Analysisreported | Terminal-Bench 2.1% of tasks | 46.8% | 80.9% |
| UGI Leaderboardreported | Requested-length error% off the requested word count | 26.0% | 15.0% |
| UGI Leaderboardreported | Style adherencescore from 0 to 1 | 0.36 | 0.38 |
| UGI Leaderboardreported | Writing scorescore out of 100 | 57.3 | 53.7 |
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.