glm-5.3
glm-5.3 has 2 tested configurations. Its highest published overall score is 70.40 at max. Inspect dimension scores, evaluation time, and reference cost below.
Overall score70.40
glm-5.3 · max
- Backend engineering
- 70
- Frontend interaction
- 78
- Knowledge & Reasoning
- 70
- Visual aesthetics
- 66
- Maintainability
- 64
- Evaluation time
- 153m 15s
- Reference cost
- $1.83
Complete evaluation · API reference cost
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 |
|---|---|---|---|---|---|---|---|---|---|
| 64.40 | 69 | 82 | 50 | 66 | 68 | 65m 34s | $2.18 | ||
| 70.40 | 70 | 78 | 70 | 66 | 64 | 153m 15s | $1.83 |
Score breakdown
- Overall score
- 70.40
- Evaluation time
- 153m 15s
- Reference cost
- $1.83
Service logic, error handling, and test verification.
Contracts and boundaries14 / 20
API contracts, input validation, and boundary behavior.
Retries and failures12 / 20
Error recovery, retry conditions, and failure handling.
CI and adversarial review14 / 20
Verification plans, risk detection, and audit evidence.
Transactions and consistency13 / 20
State transitions, concurrency control, and data consistency.
Caching and regression17 / 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
- +6.0 pts 64.40 → 70.40
- Reference cost
- −$0.35 $2.18 → $1.83
- Evaluation time
- +133.7% 65m 34s → 153m 15s
Model and effort comparisons
Data snapshot published · 10/27/2026Scoring & cost notes