Where do these coding results come from?

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.

Are reasoning settings combined into one model score?

No. Each effort pairing keeps its own recorded quality call, head-to-head record, cost and agent time. Incompatible scores are never averaged.

GLM 5.3 Flash vs GPT 5.6 Luna

GLM 5.3 Flash has a slight quality edge in 1 of 2 pairings; none is a clear lead. 2 pairings; 24 merged changes in Stet-AI/stet. Compare the individual calls below to choose which setup to investigate.

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

Compare coding models Low vs high vs xhigh/max reasoning for coding agents

Recorded setup pairings

Each row is a separate recorded comparison. Scores are not pooled or averaged across effort settings. W/D/L means wins, draws and losses; undecided judgments remain undecided.

Stet score: A setup’s average chance of beating another setup on a task, from 0 to 100; 50 is even with the field.

Setup pairing	Recorded quality call and record	Quality evidence	Cost per task	Agent time per task
GLM 5.3 Flash high vs GPT 5.6 Luna maxGLM 5.3 Flash high · high reasoningGPT 5.6 Luna max · max reasoning20/24 tasks observed by both setups.	No clear quality difference: GLM 5.3 Flash high won 3, GPT 5.6 Luna max won 2, 14 tied, 5 undecided; GLM 5.3 Flash high leads on cost and agent time: 2.0× cheaper and 1.1× faster.Head to head, GLM 5.3 Flash high won 3, tied 14 and lost 2 against GPT 5.6 Luna max, with 5 undecided.	No clear difference 3–2 head to head, 14 tiedGLM 5.3 Flash high: 47.8 (95% range 40.8–54.9)GPT 5.6 Luna max: 49.7 (95% range 44.0–55.6)	GLM 5.3 Flash high: $0.45 (from 20/24 tasks) GPT 5.6 Luna max: $0.90	GLM 5.3 Flash high: 26.6 min (from 20/24 tasks) GPT 5.6 Luna max: 28.5 min
GLM 5.3 Flash high vs GPT 5.6 Luna lowGLM 5.3 Flash high · high reasoningGPT 5.6 Luna low · codex · low reasoning20/24 tasks observed by both setups.	GLM 5.3 Flash high has a slight quality edge, 4–1 head to head with 14 tied and 5 undecided; GPT 5.6 Luna low leads on cost and agent time: 8.8× cheaper and 4.0× faster.Head to head, GLM 5.3 Flash high won 4, tied 14 and lost 1 against GPT 5.6 Luna low, with 5 undecided.	Slight edge 4–1 head to head, 14 tiedGLM 5.3 Flash high: 47.8 (95% range 40.8–54.9)GPT 5.6 Luna low: 44.6 (95% range 38.3–50.3)	GLM 5.3 Flash high: $0.45 (from 20/24 tasks) GPT 5.6 Luna low: $0.05	GLM 5.3 Flash high: 26.6 min (from 20/24 tasks) GPT 5.6 Luna low: 6.7 min

Setup pairing

Recorded quality call and record

Quality evidence

Cost per task

Agent time per task

GLM 5.3 Flash high vs GPT 5.6 Luna maxGLM 5.3 Flash high · high reasoningGPT 5.6 Luna max · max reasoning20/24 tasks observed by both setups.

No clear quality difference: GLM 5.3 Flash high won 3, GPT 5.6 Luna max won 2, 14 tied, 5 undecided; GLM 5.3 Flash high leads on cost and agent time: 2.0× cheaper and 1.1× faster.Head to head, GLM 5.3 Flash high won 3, tied 14 and lost 2 against GPT 5.6 Luna max, with 5 undecided.

No clear difference
3–2 head to head, 14 tiedGLM 5.3 Flash high: 47.8 (95% range 40.8–54.9)GPT 5.6 Luna max: 49.7 (95% range 44.0–55.6)

GLM 5.3 Flash high: $0.45 (from 20/24 tasks)

GPT 5.6 Luna max: $0.90

GLM 5.3 Flash high: 26.6 min (from 20/24 tasks)

GPT 5.6 Luna max: 28.5 min

GLM 5.3 Flash high vs GPT 5.6 Luna lowGLM 5.3 Flash high · high reasoningGPT 5.6 Luna low · codex · low reasoning20/24 tasks observed by both setups.

GLM 5.3 Flash high has a slight quality edge, 4–1 head to head with 14 tied and 5 undecided; GPT 5.6 Luna low leads on cost and agent time: 8.8× cheaper and 4.0× faster.Head to head, GLM 5.3 Flash high won 4, tied 14 and lost 1 against GPT 5.6 Luna low, with 5 undecided.

Slight edge
4–1 head to head, 14 tiedGLM 5.3 Flash high: 47.8 (95% range 40.8–54.9)GPT 5.6 Luna low: 44.6 (95% range 38.3–50.3)

GLM 5.3 Flash high: $0.45 (from 20/24 tasks)

GPT 5.6 Luna low: $0.05

GLM 5.3 Flash high: 26.6 min (from 20/24 tasks)

GPT 5.6 Luna low: 6.7 min

Questions about these results

Which is better for coding, GLM 5.3 Flash or GPT 5.6 Luna?: GLM 5.3 Flash has a slight quality edge in 1 of 2 pairings; none is a clear lead. 2 pairings; 24 merged changes in Stet-AI/stet. Compare the individual calls below to choose which setup to investigate.
Where do these coding results come from?: 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.
Are reasoning settings combined into one model score?: No. Each effort pairing keeps its own recorded quality call, head-to-head record, cost and agent time. Incompatible scores are never averaged.