DeepSeek V4.1 Flash (10/17) vs GPT 6 Luna
DeepSeek V4.1 Flash (10/17) has a slight quality edge in every pairing; 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 modelsLow 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 |
|---|---|---|---|---|
| DeepSeek V4.1 Flash (10/17) max vs GPT 6 Luna maxDeepSeek V4.1 Flash (10/17) max · codex · max reasoningGPT 6 Luna max · max reasoning22/24 tasks observed by both setups. | DeepSeek V4.1 Flash (10/17) max has a slight quality edge, 4–2 head to head with 18 tied; DeepSeek V4.1 Flash (10/17) max leads on cost and agent time: 1.3× cheaper and 1.6× faster.Head to head, DeepSeek V4.1 Flash (10/17) max won 4, tied 18 and lost 2 against GPT 6 Luna max. | Slight edge 4–2 head to head, 18 tiedDeepSeek V4.1 Flash (10/17) max: 48.3 (95% range 42.9–53.4)GPT 6 Luna max: 45.2 (95% range 39.3–51.2) | DeepSeek V4.1 Flash (10/17) max: $0.28 GPT 6 Luna max: $0.37 | DeepSeek V4.1 Flash (10/17) max: 37.2 min GPT 6 Luna max: 59.7 min |
| DeepSeek V4.1 Flash (10/17) max vs GPT 6 Luna lowDeepSeek V4.1 Flash (10/17) max · codex · max reasoningGPT 6 Luna low · low reasoning18/24 tasks observed by both setups. | DeepSeek V4.1 Flash (10/17) max has a slight quality edge, 4–1 head to head with 15 tied and 4 undecided; GPT 6 Luna low leads on cost and agent time: 5.3× cheaper and 5.4× faster.Head to head, DeepSeek V4.1 Flash (10/17) max won 4, tied 15 and lost 1 against GPT 6 Luna low, with 4 undecided. | Slight edge 4–1 head to head, 15 tiedDeepSeek V4.1 Flash (10/17) max: 48.3 (95% range 42.9–53.4)GPT 6 Luna low: 40.7 (95% range 32.0–49.2) | DeepSeek V4.1 Flash (10/17) max: $0.28 GPT 6 Luna low: $0.05 | DeepSeek V4.1 Flash (10/17) max: 37.2 min GPT 6 Luna low: 6.9 min |
Questions about these results
- Which is better for coding, DeepSeek V4.1 Flash (10/17) or GPT 6 Luna?
- DeepSeek V4.1 Flash (10/17) has a slight quality edge in every pairing; 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.