Claude Opus 5.5
Claude Opus 5.5 has 2 tested configurations. Its highest published overall score is 70.3 at max. Inspect dimension scores, evaluation time, and cost below.
Claude Opus 5.5 · max
- Backend engineering
- 49.5
- Visual aesthetics
- 44.0
- Knowledge & reasoning
- 100.0
- Frontend engineering
- 86.8
- Maintainability
- 72.0
- Evaluation time
- 203m 13s
- Cost
- $71.75
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 |
|---|---|---|---|---|---|---|---|---|---|
| 65.6 | 48.1 | 42.0 | 95.0 | 85.9 | 55.0 | 184m 43s | $62.13 | ||
| 70.3 | 49.5 | 44.0 | 100.0 | 86.8 | 72.0 | 203m 13s | $71.75 |
Score breakdown
- Overall score
- 70.3
- Evaluation time
- 203m 13s
- Cost
- $71.75
Service logic, error handling and verification.
Input/output contracts, protocol constraints and exceptional boundaries. Weight 16.69%;contribution 10.5 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 16.2 points。
Query planning, transformations, joins and result semantics. Weight 35.53%;contribution 3.2 points。
Four normalized subscores reconstruct the backend score using fixed weights; do not average them.
- B01 · 10%
- 90.0
- B02 · 10%
- 80.0
- B03 · 5%
- 80.0
- B04 · 15%
- 80.0
- B05 · 15%
- 75.0
- SQL · 25%
- 0.0
- GORM · 20%
- 26.3
What changes with higher reasoning effort
Compare overall scores, cost, and time across neighboring tested efforts.
- Overall
- +4.7 pts 65.6 → 70.3
- Cost
- +$9.62 $62.13 → $71.75
- Evaluation time
- +10.0% 184m 43s → 203m 13s