Opus 4.8 leads on quality: 19.0 points higher; GPT 6 Luna low leads on cost and agent time: 132.2× cheaper and 3.4× faster.
Stet score
GPT 6 Luna lowRank 14 of 14
21.295% range 12.5–31.2
Opus 4.8Rank 11 of 14
40.295% range 30.3–50.319.0 points higher
GPT 6 Luna low solved 0 tasks that Opus 4.8 didn’t, and Opus 4.8 solved 7 that GPT 6 Luna low didn’t. Both solved 5; neither solved 11.
stet-5276aa3fPassed testsstet-commit-bbbbae09Passed testsstet-66762b4cPassed testsstet-commit-f0caded9Equivalent to the merged fixstet-commit-5d766ddeEquivalent to the merged fixstet-43493c52Passed testsstet-6f84e978Passed testsA task counts as solved when its tests passed or a judge found the fix equivalent to the merged change, as in the leaderboard’s Solved column. Stet score also weighs code quality, so it can order the two differently. 1 of the 24 tasks is left out because one of the two has no scored result on it.
| Task | GPT 6 Luna low | Opus 4.8 | Solved by |
|---|---|---|---|
stet-3169aefa | Passed tests | Passed tests | Both |
stet-89dfbc27 | Passed tests | Passed tests | Both |
stet-b28f4fe7 | Passed tests | Passed tests | Both |
stet-232b50ba | Passed tests | Passed tests | Both |
stet-5276aa3f | Failed | Passed tests | Opus 4.8 only |
stet-commit-bbbbae09 | No change made | Passed tests | Opus 4.8 only |
stet-15439c21 | Passed tests | Passed tests | Both |
stet-66762b4c | Failed | Passed tests | Opus 4.8 only |
stet-commit-f0caded9 | Failed | Equivalent to the merged fix | Opus 4.8 only |
stet-commit-5d766dde | No change made | Equivalent to the merged fix | Opus 4.8 only |
stet-43493c52 | Failed | Passed tests | Opus 4.8 only |
stet-6f84e978 | Failed | Passed tests | Opus 4.8 only |
stet-commit-f8ae1177 | No change made | Failed | Neither |
stet-2450ca2d | Failed | Failed | Neither |
stet-commit-342d0ac0 | Failed | Failed | Neither |
stet-commit-5e075641 | Failed | Failed | Neither |
stet-commit-2826765e | Failed | Failed | Neither |
stet-commit-a8716b88 | Failed | Failed | Neither |
stet-commit-30506f8e | Failed | Failed | Neither |
stet-commit-87814583 | Failed | Failed | Neither |
stet-commit-d2ab0c25 | Failed | Failed | Neither |
stet-commit-e41c2e2d | Failed | Failed | Neither |
stet-commit-e928166f | Failed | Failed | Neither |
| Metric | GPT 6 Luna low | Opus 4.8 |
|---|---|---|
| Rank | 14 of 14 | 11 of 14 |
| Stet score | 21.295% range 12.5–31.2 | 40.295% range 30.3–50.3 |
| Solved | 5/245 passed, 0 judged equivalent | 13/2411 passed, 2 judged equivalent |
| Passed review | 5/204 not recorded | 8/231 not recorded |
| Matched merged change | 5/204 not recorded | 11/24 |
| Cost / task | $0.05 | $7.05 |
| How cost was recorded | reported by its harness | reported by its harness |
| Agent time / task | 6.9 min | 23.3 min |
| Ran in | Daytona sandbox | local Docker |
The leaderboard’s readings on groups of tasks. Each group holds fewer tasks than the overall standings, so treat these as leads to test on your repo.
11 tasks
11 tasks whose merged change was 2–149 lines in 1–2 files. The 7 small fixes below are among them.
Opus 4.8 matched the merged change more often: 9 of 11 to GPT 6 Luna low’s 5.
| GPT 6 Luna low | Opus 4.8 | |
|---|---|---|
| Matched merged change | 5/11 | 9/11 |
| Passed review | 5/11 | 8/11 |
| Cost / task | $0.02 | $1.41 |
| Agent time / task | 2.1 min | 12.8 min |
7 tasks
7 tasks whose instructions start with “Fix”. Their merged changes were 4–29 lines in 1–2 files.
Opus 4.8 matched the merged change more often: 7 of 7 to GPT 6 Luna low’s 5.
| GPT 6 Luna low | Opus 4.8 | |
|---|---|---|
| Matched merged change | 5/7 | 7/7 |
| Passed review | 5/7 | 7/7 |
| Cost / task | $0.01 | $1.17 |
| Agent time / task | 2.1 min | 9.7 min |
13 tasks
13 tasks whose merged change ran over 150 lines or more than 3 files: 113–1,420 lines in 2–19 files.
| GPT 6 Luna low | Opus 4.8 | |
|---|---|---|
| Matched merged change | 0/94 not recorded | 2/13 |
| Passed review | 0/94 not recorded | 0/121 not recorded |
| Cost / task | $0.08 | $11.83 |
| Agent time / task | 11.0 min | 32.2 min |
Stet replays your merged PRs, scores quality above pass/fail, and delivers comparison reports through your agent.
Run it on your repocompare-n24-r3b.