deepseek-v4-pro
deepseek-v4-pro · high, showing overall score. Results for all tested efforts are available below.
Overall score63.60
deepseek-v4-pro · high
- Backend engineering
- 61
- Frontend interaction
- 54
- Knowledge & Reasoning
- 75
- Visual aesthetics
- 59
- Maintainability
- 55
- Evaluation time
- 61m 4s
- Reference cost
- $0.98Partial cost
Complete evaluation · API reference cost
Backend tested · 9/21/2026Tasks & methodology
Results by reasoning effort
Swipe to see all results →
| Model / effort | Overall | Backend engineering | Frontend interaction | Knowledge & Reasoning | Visual aesthetics | Maintainability | Total eval. time | Total ref. cost | Compare |
|---|---|---|---|---|---|---|---|---|---|
| 63.60 | 61 | 54 | 75 | 59 | 55 | 61m 4s | $0.98Partial cost | ||
| 65.15 | 67 | 63 | 65 | 70 | 59 | 93m 15s | $1.15 |
Score breakdown
- Overall score
- 63.60
- Evaluation time
- 61m 4s
- Reference cost
- $0.98
Service logic, error handling, and test verification.
Contracts and boundaries10 / 20
API contracts, input validation, and boundary behavior.
Retries and failures16 / 20
Error recovery, retry conditions, and failure handling.
CI and adversarial review8 / 20
Verification plans, risk detection, and audit evidence.
Transactions and consistency13 / 20
State transitions, concurrency control, and data consistency.
Caching and regression14 / 20
Cache invalidation, data updates, and regression coverage.
What changes with higher reasoning effort
Compare overall scores, cost, and time across neighboring tested efforts.
high max
- Overall
- +1.6 pts 63.60 → 65.15
- Reference cost
- — — → $1.15
- Evaluation time
- +52.7% 61m 4s → 93m 15s
Data snapshot published · 10/27/2026Scoring & cost notes