GPT-6 Sol
GPT-6 Sol has 5 tested configurations. Its highest published overall score is 57.0 at max. Inspect dimension scores, evaluation time, and cost below.
GPT-6 Sol · max
- Backend engineering
- 46.7
- Visual aesthetics
- 25.0
- Knowledge & reasoning
- 70.0
- Frontend engineering
- 76.6
- Maintainability
- 68.0
- Evaluation time
- 149m 41s
- Cost
- $31.37
Complete evaluation · API cost
Results by reasoning effort
Swipe to see all results →
| Model / effort | Overall | Backend engineering | Visual aesthetics | Knowledge & reasoning | Frontend engineering | Maintainability | Total eval. time | Total ref. cost | Compare |
|---|---|---|---|---|---|---|---|---|---|
| 45.6 | 31.1 | 14.0 | 70.0 | 54.4 | 57.0 | 84m 19s | $20.42 | ||
| 50.0 | 37.4 | 26.0 | 65.0 | 55.5 | 69.0 | 109m 40s | $24.83 | ||
| 52.4 | 42.6 | 22.0 | 70.0 | 55.7 | 70.0 | 122m 58s | $27.66 | ||
| 56.7 | 45.3 | 30.0 | 75.0 | 71.2 | 61.0 | 141m 59s | $29.68 | ||
| 57.0 | 46.7 | 25.0 | 70.0 | 76.6 | 68.0 | 149m 41s | $31.37 |
Score breakdown
- Overall score
- 57.0
- Evaluation time
- 149m 41s
- Cost
- $31.37
Service logic, error handling and verification.
Input/output contracts, protocol constraints and exceptional boundaries. Weight 16.69%;contribution 9.7 points。
State transitions, transactions, snapshots and persistence. Weight 27.41%;contribution 19.7 points。
Dependency management, retries, caches and resource ownership. Weight 20.38%;contribution 15.7 points。
Query planning, transformations, joins and result semantics. Weight 35.53%;contribution 1.6 points。
Four normalized subscores reconstruct the backend score using fixed weights; do not average them.
- B01 · 10%
- 80.0
- B02 · 10%
- 80.0
- B03 · 5%
- 65.0
- B04 · 15%
- 80.0
- B05 · 15%
- 85.0
- SQL · 25%
- 2.3
- GORM · 20%
- 10.5
What changes with higher reasoning effort
Compare overall scores, cost, and time across neighboring tested efforts.
- Overall
- +0.3 pts 56.7 → 57.0
- Cost
- +$1.69 $29.68 → $31.37
- Evaluation time
- +5.4% 141m 59s → 149m 41s