Compare / GPT-6 Luna vs Inkling
GPT-6 Luna vs Inkling
13 results from 3 sources that measured both models. Each row is in the source's own units; there is no overall winner.
| Benchmark | Metric | GPT-6 Luna | Inkling |
|---|---|---|---|
| APEX-Agentsreported | Tasks passed% of tasks | 44.3% | 33.8% |
| Arena (formerly LMArena)reported | Confirmed task successIPS effect estimate | 0.05 | -0.22 |
| Arena (formerly LMArena)reported | Praise over complaintIPS effect estimate | -0.03 | -0.22 |
| Arena (formerly LMArena)reported | SteerabilityIPS effect estimate | 0.02 | -0.12 |
| Arena (formerly LMArena)reported | Tool groundingIPS effect estimate | 0.00 | -0.00 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,422 | 1,452 |
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,452 | 1,444 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,403 | 1,384 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,505 | 1,474 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,443 | 1,427 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,442 | 1,442 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,415 | 1,403 |
| SimpleQA Verified (Epoch AI)reported | Correct answers% of questions | 41.4% | 40.3% |
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.