Compare
A · run_d40e90vsB · run_7eb593code-reviewSwap A and B- Duration
- 37.7s -9%
- A: 41.3s
- Cost
- $0.076 -0%
- A: $0.076
- Tokens
- 22.5k -25%
- A: 29.9k
- Steps
- 11 +57%
- A: 7
- Model changed
- Claude Sonnet
- A: GPT-4.1
Steps, aligned
| Change | Step | A | B |
|---|---|---|---|
| Faster | Input policy check | 0.8s | 0.4s |
| Faster | Plan the task | 9.1s | 5.6s |
| Same | search_codebase | 3.3s | 2.7s |
| Same | Retrieve context · top 8 | 3.1s | 2.3s |
| Faster | Decide next step | 7.4s | 4.6s |
| Faster | run_tests | 4.6s | 2.6s |
| Changed | search_codebase | 10.5s | 2.2s |
| Added | Decide next step | — | 4.8s |
| Added | get_diff | — | 1.6s |
| Added | run_tests | — | 3.4s |
| Added | Write final answer | — | 5.2s |
Output diff
− 0 words + 0 words 100% unchangedReviewed 14 files. Three issues: the cache key ignores the locale, a test is skipped without a reason, and a server action is missing input validation. Tests pass. Suggested changes posted as comments, one marked blocking.