Skip to the comparison

GPT 5.6 Luna max vs SWE-2 max

GPT 5.6 Luna max leads on speed and cost; no clear quality difference. Different harnesses: GPT 5.6 Luna max ran codex, SWE-2 max ran devin. This compares setups, not models alone.

Quality
Even
No clear difference
1–9 head to head, 14 tied
Speed
2.4× faster
GPT 5.6 Luna max
28.5 min vs 68.3 min per task
Cost
3.1× cheaper
GPT 5.6 Luna max
$0.90 vs $2.76 per task; $2.76 from 23/24 tasks

What this comparison is

  • 1 real repositoryStet-AI/stet, private; not public benchmark scores
  • 24 merged changeseach run end to end by the coding agent from the repo as it was before the change
  • Graded against the merged changeby the repo’s own tests and by judges comparing each patch with it
  • Quality, cost and timea smaller sample than public benchmarks, but real agent work on a real codebase

GPT 5.6 Luna vs SWE-2: all effort pairings

24 tasksData from

Stet score

GPT 5.6 Luna maxcodex · max reasoningRank 15 of 18

45.495% range 40.0–50.2

SWE-2 maxdevin · max reasoningRank 1 of 18

62.395% range 53.4–71.6

Their 95% score ranges don’t overlap.

Quality

Head-to-head matchups

GPT 5.6 Luna max won1

Tied14

SWE-2 max won9

SWE-2 max’s win share 67%, 95% range 56–77%Ties are within judge noise.

Other quality measures

Other quality measures
MetricGPT 5.6 Luna maxSWE-2 max
Code review pass rate39%23 tasks33%24 tasks
Checklist score0.8624 tasks0.8424 tasks
Equivalence rate38%24 tasks54%24 tasks

These measures are not part of the headline Stet score and carry no confidence interval, so a small gap may be noise. The marked side has the higher recorded value, not a proven lead.

Task by task

GPT 5.6 Luna max solved 1 task that SWE-2 max didn’t, and SWE-2 max solved 7 that GPT 5.6 Luna max didn’t. Both solved 9; neither solved 7.

On the 9 tasks both solved, judges preferred SWE-2 max’s patch on 2 and called 7 even.

1Only GPT 5.6 Luna max solved

  • 4 changed lines in 1 filestet-43493c52Judges preferred GPT 5.6 Luna max’s patch

7Only SWE-2 max solved

  • 2 changed lines in 1 filestet-2450ca2dJudges preferred SWE-2 max’s patch
  • 5 changed lines in 2 filesstet-15439c21Judges preferred SWE-2 max’s patch
  • 23 changed lines in 1 filestet-6f84e978Judges preferred SWE-2 max’s patch
  • 481 changed lines in 3 filesstet-commit-5e075641Judges preferred SWE-2 max’s patch
  • 552 changed lines in 11 filesstet-commit-60a0e6c2Judges preferred SWE-2 max’s patch
  • 564 changed lines in 10 filesstet-commit-a8716b88Judges preferred SWE-2 max’s patch
  • 640 changed lines in 8 filesstet-commit-342d0ac0Judges preferred SWE-2 max’s patch

9Both solved

  • 6 changed lines in 1 filestet-3169aefa
  • 9 changed lines in 1 filestet-89dfbc27
  • 10 changed lines in 1 filestet-232b50ba
  • 23 changed lines in 2 filesstet-5276aa3f
  • 29 changed lines in 1 filestet-b28f4fe7
  • 84 changed lines in 1 filestet-66762b4cJudges preferred SWE-2 max’s patch
  • 113 changed lines in 4 filesstet-commit-f0caded9Judges preferred SWE-2 max’s patch
  • 218 changed lines in 2 filesstet-commit-5d766dde
  • 941 changed lines in 7 filesstet-commit-bbbbae09

7Neither solved

  • 149 changed lines in 2 filesstet-commit-e928166f
  • 476 changed lines in 17 filesstet-commit-30506f8e
  • 763 changed lines in 8 filesstet-commit-87814583
  • 977 changed lines in 5 filesstet-commit-e41c2e2d
  • 1,236 changed lines in 16 filesstet-commit-d2ab0c25
  • 1,312 changed lines in 9 filesstet-commit-f8ae1177
  • 1,420 changed lines in 19 filesstet-commit-2826765e

Bar: changed lines in the merged change, log scale

A 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.

All 24 compared tasks
Task by task
TaskGPT 5.6 Luna maxSWE-2 maxSolved byBetter patch
stet-89dfbc27 · 9 changed lines in 1 filePassed tests$0.25 · 6.8 minPassed tests$0.71 · 4.5 minBothEven
stet-b28f4fe7 · 29 changed lines in 1 filePassed tests$0.54 · 12.2 minPassed tests$2.02 · 16.7 minBothEven
stet-232b50ba · 10 changed lines in 1 filePassed tests$0.26 · 16.0 minPassed tests$1.56 · 11.9 minBothEven
stet-3169aefa · 6 changed lines in 1 filePassed tests$0.10 · 3.8 minPassed tests$1.03 · 5.3 minBothEven
stet-5276aa3f · 23 changed lines in 2 filesPassed tests$0.18 · 5.8 minPassed tests$1.03 · 10.7 minBothEven
stet-66762b4c · 84 changed lines in 1 fileEquivalent to the merged fix$0.08 · 4.9 minPassed tests$1.77 · 11.2 minBothSWE-2 max
stet-commit-bbbbae09 · 941 changed lines in 7 filesPassed tests$1.21 · 41.6 minPassed tests$0.61 · 89.7 minBothEven
stet-commit-5d766dde · 218 changed lines in 2 filesEquivalent to the merged fix$0.48 · 19.3 minEquivalent to the merged fix$4.04 · 20.0 minBothEven
stet-43493c52 · 4 changed lines in 1 filePassed tests$0.14 · 7.3 minFailed$3.52 · 35.3 minGPT 5.6 Luna max onlyGPT 5.6 Luna max
stet-commit-f0caded9 · 113 changed lines in 4 filesEquivalent to the merged fix$0.90 · 25.2 minPassed tests$3.91 · 22.6 minBothSWE-2 max
stet-15439c21 · 5 changed lines in 2 filesFailed$0.14 · 17.9 minPassed tests$2.24 · 34.3 minSWE-2 max onlySWE-2 max
stet-commit-f8ae1177 · 1,312 changed lines in 9 filesFailed$1.38 · 52.9 minFailed300.0 minNeitherEven
stet-2450ca2d · 2 changed lines in 1 fileFailed$0.09 · 9.1 minPassed tests$1.18 · 11.8 minSWE-2 max onlySWE-2 max
stet-commit-342d0ac0 · 640 changed lines in 8 filesFailed$2.15 · 66.8 minEquivalent to the merged fix$2.39 · 146.4 minSWE-2 max onlySWE-2 max
stet-6f84e978 · 23 changed lines in 1 fileFailed$0.55 · 14.3 minPassed tests$2.06 · 42.4 minSWE-2 max onlySWE-2 max
stet-commit-5e075641 · 481 changed lines in 3 filesFailed$0.99 · 34.4 minPassed tests$2.64 · 86.0 minSWE-2 max onlySWE-2 max
stet-commit-60a0e6c2 · 552 changed lines in 11 filesFailed$0.84 · 26.4 minPassed tests$0.66 · 106.1 minSWE-2 max onlySWE-2 max
stet-commit-2826765e · 1,420 changed lines in 19 filesFailed$1.52 · 57.1 minFailed$7.00 · 144.5 minNeitherEven
stet-commit-a8716b88 · 564 changed lines in 10 filesFailed$2.65 · 60.3 minEquivalent to the merged fix$1.29 · 102.3 minSWE-2 max onlySWE-2 max
stet-commit-e928166f · 149 changed lines in 2 filesFailed$1.25 · 37.4 minFailed$1.49 · 59.4 minNeitherEven
stet-commit-30506f8e · 476 changed lines in 17 filesFailed$2.47 · 54.8 minFailed$5.49 · 164.8 minNeitherEven
stet-commit-87814583 · 763 changed lines in 8 filesFailed$1.17 · 40.8 minFailed$5.19 · 73.6 minNeitherEven
stet-commit-d2ab0c25 · 1,236 changed lines in 16 filesFailed$1.27 · 36.4 minFailed$5.66 · 68.9 minNeitherEven
stet-commit-e41c2e2d · 977 changed lines in 5 filesFailed$0.93 · 32.4 minFailed$6.08 · 69.8 minNeitherEven

Cost and time on each task

Each dot is one task, on a log scale: filled where the setup solved it, hollow where it didn’t. The line marks the median.

Cost per task

GPT 5.6 Luna maxmean $0.90 · median $0.87

SWE-2 maxmean $2.76 · median $2.06from 23/24 tasks

GPT 5.6 Luna max cost less on 20 of 23 tasks; SWE-2 max on 3.

Agent time per task

GPT 5.6 Luna maxmean 28.5 min · median 25.8 min

SWE-2 maxmean 68.3 min · median 50.9 min

GPT 5.6 Luna max finished faster on 21 of 24 tasks; SWE-2 max on 3.

Where each setup is the better choice

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.

Small fixes

10 tasks · 2–84 changed lines in 1–2 files

SWE-2 max solved 2 more, 9 of 10 to 7; GPT 5.6 Luna max cost 7.4× less per task.

GPT 5.6 Luna max was 1.9× faster.

Small fixes
GPT 5.6 Luna maxSWE-2 max
Solved7/109/10
Cost / task$0.23$1.71
Agent time / task9.8 min18.4 min

Large, cross-cutting changes

14 tasks · 113–1,420 changed lines in up to 19 files

SWE-2 max solved 7 of 14 to GPT 5.6 Luna max’s 3.

GPT 5.6 Luna max was 2.5× faster.

Large, cross-cutting changes
GPT 5.6 Luna maxSWE-2 max
Solved3/147/14
Cost / task$1.37$3.57from 13/14 tasks
Agent time / task41.9 min103.9 min

Side by side

Side by side
MetricGPT 5.6 Luna maxSWE-2 max
Harnesscodexdevin
Reasoning effortmaxmax
Rank15 of 181 of 18
Stet score45.495% range 40.0–50.262.395% range 53.4–71.6
Solved10/247 passed, 3 judged equivalent16/2413 passed, 3 judged equivalent
Cost / task$0.90$2.76from 23/24 tasks
Agent time / task28.5 min68.3 min

Questions

Which is better for coding, GPT 5.6 Luna max or SWE-2 max?
No clear quality difference: SWE-2 max won 9, GPT 5.6 Luna max won 1, 14 tied; GPT 5.6 Luna max leads on cost and agent time: 3.1× cheaper and 2.4× faster. Measured on 24 merged changes from Stet-AI/stet.
How were GPT 5.6 Luna max and SWE-2 max tested?
Both coding agents worked the same 24 merged changes from the private Stet-AI/stet repository, each in the real repo as it was before the change. The repo’s tests and judges graded every patch against the merged change; no public benchmark scores are used. Stet score: A setup’s average chance of beating another setup on a task, from 0 to 100; 50 is even with the field.

See which setup wins on your repo

Stet replays your merged changes, scores quality above pass/fail, and delivers comparison reports through your agent.

Get started

About this data

  • 24 tasks, each a merged change from the private Stet-AI/stet repository replayed from the repo as it was just before the change.
  • Data from .
  • Both setups come from the same leaderboard run of 18 setups: bmk_d4ba310fdf50db4ccb2092e3.