# GPT 6 Luna low vs Pareto 26.9 medium: real-repo coding test

Stet publishes a public search API over its indexed pages; the guide is at https://www.stet.sh/agents/agent-access.

Source: https://www.stet.sh/leaderboard/compare/gpt-6-luna-low-vs-pareto-26-9-medium
Source observed: 2026-11-19T11:08:37.634Z
Last checked: 2026-11-19T11:08:37.634+00:00
Indexed: 2026-11-19T11:18:21.693+00:00

---

# GPT 6 Luna low vs Pareto 26.9 medium

We compare two coding-agent setups on real repository work. GPT 6 Luna low leads on speed and cost; we found no clear quality difference. GPT 6 Luna low ran `codex` with low reasoning, while Pareto 26.9 medium ran `claude-code` with medium reasoning, so this compares setups rather than models alone.

## At a glance

- **Quality:** Even; no clear difference. GPT 6 Luna low won 1 head-to-head task, Pareto 26.9 medium won 5, 14 were tied, and 4 were undecided.
- **Speed:** GPT 6 Luna low was **5.7× faster**: 6.9 minutes versus 39.2 minutes per task.
- **Cost:** GPT 6 Luna low was **112× cheaper**: $0.05 versus $5.98 per task.

## What this comparison is

- **1 real repository:** private `Stet-AI/stet`; these are not public benchmark scores.
- **24 merged changes:** each setup ran end to end from the repository state before the change.
- **Grading against the merged change:** we used the repository’s tests and judges comparing each patch with the merged change.
- **Quality, cost, and time:** this is a smaller sample than public benchmarks, but it measures agent work on a real codebase.

[GPT 6 Luna vs Pareto 26.9: all effort pairings](https://www.stet.sh/leaderboard/compare/gpt-6-luna-vs-pareto-26-9)

**24 tasks · Data from November 19, 2026**

## Stet score

| Setup | Harness and reasoning | Rank | Stet score | 95% range |
| --- | --- | --- | --- | --- |
| GPT 6 Luna low | `codex` · low reasoning | 18 of 18 | 40.5 | 31.7–49.2 |
| Pareto 26.9 medium | `claude-code` · medium reasoning | 7 of 18 | 52.3 | 44.9–59.5 |

Their 95% score ranges overlap. Stet score is a setup’s average chance of beating another setup on a task, from 0 to 100; 50 is even with the field.

## Quality

### Head-to-head matchups

- GPT 6 Luna low won **1**.
- **14** were tied.
- **4** were undecided.
- Pareto 26.9 medium won **5**.
- Pareto 26.9 medium’s win share was **61%**, with a 95% range of **50–72%**.

Ties are within judge noise.

### Other quality measures

| Metric | GPT 6 Luna low | Pareto 26.9 medium |
| --- | --- | --- |
| Code review pass rate | 21% — 24 tasks | 35% — 23 tasks |
| Checklist score | 0.51 — 24 tasks | 0.80 — 24 tasks |
| Equivalence rate | 21% — 24 tasks | 39% — 23 tasks |

These measures are not part of the headline Stet score and have no confidence interval. A small gap may be noise; the higher recorded value is not a proven lead.

## Task by task

GPT 6 Luna low solved 1 task that Pareto 26.9 medium did not, and Pareto 26.9 medium solved 5 that GPT 6 Luna low did not. Both solved 4; neither solved 10. On the 4 tasks both solved, judges called every patch even.

### Only GPT 6 Luna low solved — 1

- 5 changed lines in 2 files — `stet-15439c21`; judges preferred GPT 6 Luna low’s patch.

### Only Pareto 26.9 medium solved — 5

- 2 changed lines in 1 file — `stet-2450ca2d`; judges preferred Pareto 26.9 medium’s patch.
- 4 changed lines in 1 file — `stet-43493c52`; judges preferred Pareto 26.9 medium’s patch.
- 23 changed lines in 2 files — `stet-5276aa3f`; judges preferred Pareto 26.9 medium’s patch.
- 84 changed lines in 1 file — `stet-66762b4c`; judges preferred Pareto 26.9 medium’s patch.
- 113 changed lines in 4 files — `stet-commit-f0caded9`; judges preferred Pareto 26.9 medium’s patch.

### Both solved — 4

- 6 changed lines in 1 file — `stet-3169aefa`.
- 9 changed lines in 1 file — `stet-89dfbc27`.
- 10 changed lines in 1 file — `stet-232b50ba`.
- 29 changed lines in 1 file — `stet-b28f4fe7`.

### Neither solved — 10

- 23 changed lines in 1 file — `stet-6f84e978`.
- 149 changed lines in 2 files — `stet-commit-e928166f`.
- 476 changed lines in 17 files — `stet-commit-30506f8e`.
- 481 changed lines in 3 files — `stet-commit-5e075641`.
- 564 changed lines in 10 files — `stet-commit-a8716b88`.
- 640 changed lines in 8 files — `stet-commit-342d0ac0`.
- 763 changed lines in 8 files — `stet-commit-87814583`.
- 977 changed lines in 5 files — `stet-commit-e41c2e2d`.
- 1,236 changed lines in 16 files — `stet-commit-d2ab0c25`.
- 1,420 changed lines in 19 files — `stet-commit-2826765e`.

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 setups differently. Four of the 24 tasks are left out because one setup has no scored result for them.

## All 20 compared tasks

| Task | GPT 6 Luna low | Pareto 26.9 medium | Solved by | Better patch |
| --- | --- | --- | --- | --- |
| `stet-89dfbc27` · 9 changed lines in 1 file | Passed tests · $0.01 · 3.3 min | Passed tests · $3.94 · 47.2 min | Both | Even |
| `stet-b28f4fe7` · 29 changed lines in 1 file | Passed tests · $0.01 · 1.6 min | Passed tests · $3.17 · 22.7 min | Both | Even |
| `stet-232b50ba` · 10 changed lines in 1 file | Passed tests · $0.01 · 1.8 min | Passed tests · $3.09 · 42.8 min | Both | Even |
| `stet-3169aefa` · 6 changed lines in 1 file | Passed tests · $0.01 · 1.7 min | Passed tests · $2.42 · 14.8 min | Both | Even |
| `stet-5276aa3f` · 23 changed lines in 2 files | Failed · $0.01 · 2.6 min | Passed tests · $0.69 · 13.0 min | Pareto 26.9 medium only | Pareto 26.9 medium |
| `stet-66762b4c` · 84 changed lines in 1 file | Failed · $0.01 · 1.5 min | Passed tests · $3.39 · 50.6 min | Pareto 26.9 medium only | Pareto 26.9 medium |
| `stet-43493c52` · 4 changed lines in 1 file | Failed · $0.01 · 1.7 min | Passed tests · $2.04 · 29.5 min | Pareto 26.9 medium only | Pareto 26.9 medium |
| `stet-commit-f0caded9` · 113 changed lines in 4 files | Failed · $0.03 · 4.5 min | Equivalent to the merged fix · $8.97 · 60.8 min | Pareto 26.9 medium only | Pareto 26.9 medium |
| `stet-15439c21` · 5 changed lines in 2 files | Passed tests · $0.02 · 3.1 min | Failed · $2.73 · 36.0 min | GPT 6 Luna low only | GPT 6 Luna low |
| `stet-2450ca2d` · 2 changed lines in 1 file | Failed · $0.01 · 1.1 min | Passed tests · $0.61 · 12.3 min | Pareto 26.9 medium only | Pareto 26.9 medium |
| `stet-commit-342d0ac0` · 640 changed lines in 8 files | Failed · $0.31 · 40.1 min | Failed · $7.95 · 42.6 min | Neither | Even |
| `stet-6f84e978` · 23 changed lines in 1 file | Failed · $0.01 · 1.3 min | Failed · $2.15 · 20.9 min | Neither | Even |
| `stet-commit-5e075641` · 481 changed lines in 3 files | Failed · $0.01 · 1.8 min | Failed · $2.54 · 23.6 min | Neither | Even |
| `stet-commit-2826765e` · 1,420 changed lines in 19 files | Failed · $0.02 · 4.2 min | Failed · $8.80 · 40.3 min | Neither | Even |
| `stet-commit-a8716b88` · 564 changed lines in 10 files | Failed · $0.21 · 27.1 min | Failed · $13.86 · 42.7 min | Neither | Even |
| `stet-commit-e928166f` · 149 changed lines in 2 files | Failed · $0.06 · 3.1 min | Failed · $7.09 · 34.0 min | Neither | Even |
| `stet-commit-30506f8e` · 476 changed lines in 17 files | Failed · $0.23 · 38.4 min | Failed · $16.87 · 66.0 min | Neither | Even |
| `stet-commit-87814583` · 763 changed lines in 8 files | Failed · $0.02 · 1.0 min | Failed · $11.99 · 89.0 min | Neither | Even |
| `stet-commit-d2ab0c25` · 1,236 changed lines in 16 files | Failed · $0.10 · 7.2 min | Failed · $1.11 · 8.3 min | Neither | Even |
| `stet-commit-e41c2e2d` · 977 changed lines in 5 files | Failed · $0.09 · 12.1 min | Failed · $9.40 · 67.9 min | Neither | Even |

## Cost and time on each task

Each dot represents one task on a log scale. Filled dots indicate a solved task; hollow dots indicate an unsolved task. The line marks the median.

- **Cost per task:** GPT 6 Luna low — mean $0.06, median $0.02. Pareto 26.9 medium — mean $5.64, median $3.28. GPT 6 Luna low cost less on all 20 tasks.
- **Agent time per task:** GPT 6 Luna low — mean 8.0 min, median 2.8 min. Pareto 26.9 medium — mean 38.3 min, median 38.2 min. GPT 6 Luna low finished faster on all 20 tasks.

## Where each setup is the better choice

These are readings on smaller task groups, so use them as leads to test on your repository.

### Small fixes

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

Pareto 26.9 medium solved 3 more: 8 of 10 versus 5 of 10. GPT 6 Luna low cost 204× less per task and was 15× faster.

| Metric | GPT 6 Luna low | Pareto 26.9 medium |
| --- | --- | --- |
| Solved | 5/10 | 8/10 |
| Cost / task | $0.01 | $2.42 |
| Agent time / task | 2.0 min | 29.0 min |

### Large, cross-cutting changes

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

Pareto 26.9 medium solved 1 more: 1 of 10 versus 0 of 10. GPT 6 Luna low cost 82× less per task and was 3.4× faster.

| Metric | GPT 6 Luna low | Pareto 26.9 medium |
| --- | --- | --- |
| Solved | 0/10 | 1/10 |
| Cost / task | $0.11 | $8.86 |
| Agent time / task | 13.9 min | 47.5 min |

## Side by side

| Metric | GPT 6 Luna low | Pareto 26.9 medium |
| --- | --- | --- |
| Harness | `codex` | `claude-code` |
| Reasoning effort | low | medium |
| Rank | 18 of 18 | 7 of 18 |
| Stet score | 40.5; 95% range 31.7–49.2 | 52.3; 95% range 44.9–59.5 |
| Solved | 5/24 — 5 passed, 0 judged equivalent | 11/24 — 10 passed, 1 judged equivalent |
| Cost / task | $0.05 | $5.98 |
| Agent time / task | 6.9 min | 39.2 min |

## Research on these models

- [Cost per successful task ranked the coding agent that lost every head to head 1st](https://www.stet.sh/blog/cost-per-successful-task)
- [A benchmark mined from my own repo agreed with FrontierCode](https://www.stet.sh/blog/benchmark-agreed-with-frontier-code)
- [I ran GPT 6.1 Sol against GPT 6 Sol on 24 tasks from my own repo, and found that it maintained quality while cutting cost in half](https://www.stet.sh/blog/gpt-6-1-sol-vs-gpt-6-sol)

## Questions

### Which is better for coding, GPT 6 Luna low or Pareto 26.9 medium?

We found no clear quality difference: Pareto 26.9 medium won 5, GPT 6 Luna low won 1, 14 were tied, and 4 were undecided. GPT 6 Luna low led on cost and agent time: 112× cheaper and 5.7× faster. We measured this on 24 merged changes from `Stet-AI/stet`.

### How were GPT 6 Luna low and Pareto 26.9 medium tested?

Both coding agents worked on the same 24 merged changes from our private `Stet-AI/stet` repository, each in the real repository state before the change. The repository’s tests and judges graded every patch against the merged change; we used no public benchmark scores.

## See which setup wins on your repository

We replay your merged changes, score quality above pass/fail, and deliver comparison reports through your agent.

[Get started](https://www.stet.sh/private)

## About this data

- 24 tasks, each a merged change from the private `Stet-AI/stet` repository, replayed from the repository state just before the change.
- Data from November 19, 2026.
- Both setups came from the same leaderboard run of 18 setups: `bmk_d4ba310fdf50db4ccb2092e3`.

[See all 18 setups on the leaderboard](https://www.stet.sh/leaderboard) · [How Stet measures](https://www.stet.sh/methodology)

## More comparisons

- [GPT 6 Luna low vs GPT 5.6 Luna low](https://www.stet.sh/leaderboard/compare/gpt-5-6-luna-low-vs-gpt-6-luna-low)
- [GPT 6 Luna low vs Haiku 5.5 low](https://www.stet.sh/leaderboard/compare/gpt-6-luna-low-vs-haiku-5-5-low)
- [GPT 6 Luna low vs GPT 6 Luna max](https://www.stet.sh/leaderboard/compare/gpt-6-luna-low-vs-gpt-6-luna-max)
- [GPT 6 Luna low vs GPT 5.6 Luna max](https://www.stet.sh/leaderboard/compare/gpt-5-6-luna-max-vs-gpt-6-luna-low)
- [GPT 6 Luna low vs Opus 4.8 medium](https://www.stet.sh/leaderboard/compare/gpt-6-luna-low-vs-opus-4-8-medium)
- [Pareto 26.9 medium vs Grok 4.7 high](https://www.stet.sh/leaderboard/compare/grok-4-7-high-vs-pareto-26-9-medium)
- [Pareto 26.9 medium vs DeepSeek V4.1 Flash max](https://www.stet.sh/leaderboard/compare/deepseek-v4-1-flash-max-vs-pareto-26-9-medium)
- [Pareto 26.9 medium vs GPT 6 Sol medium](https://www.stet.sh/leaderboard/compare/gpt-6-sol-medium-vs-pareto-26-9-medium)
- [Pareto 26.9 medium vs DeepSeek V4.1 Flash (10/17) max](https://www.stet.sh/leaderboard/compare/deepseek-v4-1-flash-10-17-max-vs-pareto-26-9-medium)
- [Pareto 26.9 medium vs GPT 6.1 Sol medium](https://www.stet.sh/leaderboard/compare/gpt-6-1-sol-medium-vs-pareto-26-9-medium)
- [All 153 comparisons](https://www.stet.sh/leaderboard/compare)

---

[Read this page’s summary](https://www.stet.sh/agents/api/public/v1/page-tldr?url=https%3A%2F%2Fwww.stet.sh%2Fleaderboard%2Fcompare%2Fgpt-6-luna-low-vs-pareto-26-9-medium&trace=t_5da45423e484)

## Ask a follow-up about this page

Need a more specific answer? Ask the exact question you have. It returns a fresh, cited answer from Stet’s knowledge base. No account or API key required.

Open this URL with your own question, URL-encoded. Include your question in q. All research inputs are optional; no account or setup.

`GET https://www.stet.sh/agents/agent/faq?q={question}&trace=t_5da45423e484`

[Tool guide and examples](https://www.stet.sh/agents/agent-access) · [OpenAPI](https://www.stet.sh/agents/api/public/v1/openapi.json) · [Page TLDR form](https://www.stet.sh/agents/agent-tools/page-tldr?url=https%3A%2F%2Fwww.stet.sh%2Fleaderboard%2Fcompare%2Fgpt-6-luna-low-vs-pareto-26-9-medium). Opening a form does not execute a tool.