gpt-6-astra
gpt-6-astra · high, showing overall score. Results for all tested efforts are available below.
Overall score84.85
gpt-6-astra · high
- Backend engineering
- 92
- Frontend interaction
- 96
- Knowledge & Reasoning
- 85
- Visual aesthetics
- 70
- Maintainability
- 70
- Evaluation time
- 79m 52s
- Reference cost
- $4.40
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 |
|---|---|---|---|---|---|---|---|---|---|
| 85.25 | 90 | 94 | 85 | 78 | 70 | 61m 27s | $3.70 | ||
| 84.00 | 91 | 89 | 85 | 75 | 70 | 72m 41s | $3.76 | ||
| 84.85 | 92 | 96 | 85 | 70 | 70 | 79m 52s | $4.40 | ||
| 88.80 | 95 | 96 | 95 | 75 | 61 | 83m 56s | $7.01 | ||
| 88.50 | 97 | 94 | 95 | 73 | 61 | 93m 26s | $8.35 |
Score breakdown
- Overall score
- 84.85
- Evaluation time
- 79m 52s
- Reference cost
- $4.40
Service logic, error handling, and test verification.
Contracts and boundaries19 / 20
API contracts, input validation, and boundary behavior.
Retries and failures19 / 20
Error recovery, retry conditions, and failure handling.
CI and adversarial review16 / 20
Verification plans, risk detection, and audit evidence.
Transactions and consistency18 / 20
State transitions, concurrency control, and data consistency.
Caching and regression20 / 20
Cache invalidation, data updates, and regression coverage.
What changes with higher reasoning effort
Compare overall scores, cost, and time across neighboring tested efforts.
medium high
- Overall
- +0.8 pts 84.00 → 84.85
- Reference cost
- +$0.64 $3.76 → $4.40
- Evaluation time
- +9.9% 72m 41s → 79m 52s
Data snapshot published · 10/27/2026Scoring & cost notes