Model comparison
Compare models or reasoning efforts by score, evaluation time and estimated cost.
Available modelsSelected 2 / 3
Measured results
Differences are relative to the first column
gpt-6-lunaReasoning −20 pts · Aesthetics +11 pts
| Metric | k2.8 preview · high | gpt-6-luna · xhigh |
|---|---|---|
| Overall | 63.55 | 62.00−1.55 pts |
| Backend engineering | 54 | 62+8 pts |
| Frontend interaction | 72 | 80+8 pts |
| Knowledge & reasoning | 70 | 50−20 pts |
| Visual aesthetics | 57 | 68+11 pts |
| Maintainability | 53 | 59+6 pts |
Cost & time
Complete evaluation| Metric | k2.8 preview · high | gpt-6-luna · xhigh |
|---|---|---|
| Reference cost | $0.66 | $0.061−$0.60 |
| Evaluation time | 70m 58s | 63m 52s−10% |
Dimension breakdown
Expand to inspect each subaxis.
Backend engineeringSubscores
| Scoring dimension | k2.8 previewhigh | gpt-6-lunaxhigh |
|---|---|---|
| Contracts and boundaries | 9 / 20 | 10 / 20 |
| Retries and failures | 15 / 20 | 16 / 20 |
| CI and adversarial review | 14 / 20 | 9 / 20 |
| Transactions and consistency | 9 / 20 | 13 / 20 |
| Caching and regression | 7 / 20 | 14 / 20 |
Frontend interactionSubscores
| Scoring dimension | k2.8 previewhigh | gpt-6-lunaxhigh |
|---|---|---|
| Layout and adaptation | 16 / 20 | 18 / 20 |
| Interaction behavior | 18 / 20 | 20 / 20 |
| State continuity | 15 / 20 | 15 / 20 |
| Accessibility and feedback | 10 / 15 | 10 / 15 |
| State validation and regression | 13 / 25 | 17 / 25 |
Knowledge & reasoningSubscores
| Scoring dimension | k2.8 previewhigh | gpt-6-lunaxhigh |
|---|---|---|
| Mathematical reasoning | 5 / 15 | 0 / 15 |
| Computer science & AI | 5 / 15 | 5 / 15 |
| Science and engineering | 5 / 10 | 10 / 10 |
| Life sciences and medicine | 20 / 20 | 15 / 20 |
| Humanities and general knowledge | 35 / 40 | 20 / 40 |
Visual aestheticsSubscores
| Scoring dimension | k2.8 previewhigh | gpt-6-lunaxhigh |
|---|---|---|
| Composition | 16 / 20 | 19 / 20 |
| Spacing and alignment | 17 / 20 | 16 / 20 |
| Typography and expression | 16 / 20 | 17 / 20 |
| Color and surfaces | 19 / 20 | 19 / 20 |
| Control styling | 15 / 20 | 17 / 20 |
Subscores retain their original scale; the axis score uses a consistent conversion.
MaintainabilitySubscores
| Scoring dimension | k2.8 previewhigh | gpt-6-lunaxhigh |
|---|---|---|
| Semantic clarity | 14 / 20 | 15 / 20 |
| Control-flow organization | 19 / 20 | 20 / 20 |
| Locality of changes | 15 / 20 | 15 / 20 |
| State and lifecycle | 17 / 20 | 18 / 20 |
| Implementation economy | 16 / 20 | 16 / 20 |
Subscores retain their original scale; the axis score uses a consistent conversion.