# DeepSeek V4.1 Flash max vs GPT 6.1 Sol medium for coding

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/deepseek-v4-1-flash-max-vs-gpt-6-1-sol-medium
Source observed: 2026-11-19T11:08:38.832Z
Last checked: 2026-11-19T11:08:38.832+00:00
Indexed: 2026-11-19T11:19:15.091+00:00

---

# DeepSeek V4.1 Flash max vs GPT 6.1 Sol medium

GPT 6.1 Sol medium leads on speed; DeepSeek V4.1 Flash max leads on cost; we found no clear quality difference. This compares setups, not models alone: DeepSeek ran `deepseek-harness` with max reasoning, while GPT ran `codex` with medium reasoning.

## Headline results

- **Quality:** Even. DeepSeek won 1 head-to-head task, GPT won 4, 13 were tied, and 6 were undecided.
- **Speed:** GPT was **4.2× faster**: 11.8 versus 49.3 minutes per task.
- **Cost:** DeepSeek was **2.2× cheaper**: $0.49 versus $1.10 per task. The $0.49 figure is from 23 of 24 tasks.

## Test design and data

We ran both setups end to end on 24 merged changes from the private `Stet-AI/stet` repository, starting from the state before each change. Repository tests and judges compared each patch with the merged change. Data is from **November 19, 2026**; these are real-repository results, not public benchmark scores.

## Scores and quality

| Setup | Harness and reasoning | Rank | Stet score | 95% range |
|---|---|---:|---:|---:|
| DeepSeek V4.1 Flash max | `deepseek-harness`, max | 8 of 18 | 51.0 | 42.1–60.1 |
| GPT 6.1 Sol medium | `codex`, medium | 4 of 18 | 54.7 | 46.0–63.4 |

Their 95% score ranges overlap. GPT’s head-to-head win share was 57%, with a 95% range of 45–69%; ties are within judge noise.

### Other quality measures

| Metric | DeepSeek V4.1 Flash max | GPT 6.1 Sol medium |
|---|---:|---:|
| Code review pass rate | 33% — 24 tasks | 43% — 23 tasks |
| Checklist score | 0.83 — 24 tasks | 0.88 — 24 tasks |
| Equivalence rate | 43% — 23 tasks | 43% — 23 tasks |

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

## Task-by-task outcome

DeepSeek solved 1 task that GPT did not. GPT solved 0 that DeepSeek did not. Both solved 9; neither solved 8. On the 9 tasks both solved, judges preferred GPT’s patch on 3 and called 6 even.

### Only DeepSeek V4.1 Flash max solved

- 5 changed lines in 2 files — `stet-15439c21`; judges preferred DeepSeek’s patch.

### Only GPT 6.1 Sol medium solved

None.

### Both solved

- 4 changed lines in 1 file — `stet-43493c52`; judges preferred GPT’s patch.
- 9 changed lines in 1 file — `stet-89dfbc27`.
- 10 changed lines in 1 file — `stet-232b50ba`.
- 23 changed lines in 2 files — `stet-5276aa3f`.
- 29 changed lines in 1 file — `stet-b28f4fe7`.
- 84 changed lines in 1 file — `stet-66762b4c`.
- 218 changed lines in 2 files — `stet-commit-5d766dde`.
- 640 changed lines in 8 files — `stet-commit-342d0ac0`; judges preferred GPT’s patch.
- 1,312 changed lines in 9 files — `stet-commit-f8ae1177`; judges preferred GPT’s patch.

### Neither solved

- 2 changed lines in 1 file — `stet-2450ca2d`.
- 23 changed lines in 1 file — `stet-6f84e978`.
- 113 changed lines in 4 files — `stet-commit-f0caded9`.
- 149 changed lines in 2 files — `stet-commit-e928166f`; judges preferred GPT’s patch.
- 476 changed lines in 17 files — `stet-commit-30506f8e`.
- 552 changed lines in 11 files — `stet-commit-60a0e6c2`.
- 763 changed lines in 8 files — `stet-commit-87814583`.
- 1,420 changed lines in 19 files — `stet-commit-2826765e`.

Six of the 24 tasks are left out of the compared-task view because one setup has no scored result on them.

## All 18 compared tasks

| Task | DeepSeek V4.1 Flash max | GPT 6.1 Sol medium | Solved by | Better patch |
|---|---|---|---|---|
| `stet-89dfbc27` · 9 lines, 1 file | Passed · $0.09 · 9.3 min | Passed · $0.18 · 2.8 min | Both | Even |
| `stet-b28f4fe7` · 29 lines, 1 file | Passed · $0.22 · 21.5 min | Passed · $0.73 · 7.1 min | Both | Even |
| `stet-232b50ba` · 10 lines, 1 file | Passed · $0.34 · 59.2 min | Passed · $0.24 · 3.8 min | Both | Even |
| `stet-5276aa3f` · 23 lines, 2 files | Passed · $0.16 · 37.9 min | Passed · $0.54 · 6.5 min | Both | Even |
| `stet-66762b4c` · 84 lines, 1 file | Passed · $0.09 · 15.1 min | Passed · $0.32 · 3.6 min | Both | Even |
| `stet-commit-5d766dde` · 218 lines, 2 files | Equivalent · $0.27 · 35.4 min | Equivalent · $0.77 · 9.8 min | Both | Even |
| `stet-43493c52` · 4 lines, 1 file | Passed · $0.22 · 63.4 min | Passed · $0.39 · 5.7 min | Both | GPT |
| `stet-commit-f0caded9` · 113 lines, 4 files | Failed · $0.25 · 42.5 min | Failed · $0.87 · 9.2 min | Neither | Even |
| `stet-15439c21` · 5 lines, 2 files | Passed · $0.15 · 45.6 min | Failed · $0.61 · 8.6 min | DeepSeek only | DeepSeek |
| `stet-commit-f8ae1177` · 1,312 lines, 9 files | Passed · $1.72 · 113.6 min | Passed · $1.83 · 28.7 min | Both | GPT |
| `stet-2450ca2d` · 2 lines, 1 file | Failed · $0.15 · 29.5 min | Failed · $0.26 · 4.2 min | Neither | Even |
| `stet-commit-342d0ac0` · 640 lines, 8 files | Equivalent · $1.05 · 59.4 min | Equivalent · $1.49 · 14.4 min | Both | GPT |
| `stet-6f84e978` · 23 lines, 1 file | Failed · $0.24 · 28.7 min | Failed · $0.50 · 5.4 min | Neither | Even |
| `stet-commit-60a0e6c2` · 552 lines, 11 files | Failed · $0.58 · 47.5 min | Failed · $1.38 · 14.2 min | Neither | Even |
| `stet-commit-2826765e` · 1,420 lines, 19 files | Failed · $0.80 · 70.4 min | Failed · $1.60 · 13.8 min | Neither | Even |
| `stet-commit-e928166f` · 149 lines, 2 files | Failed · $0.58 · 58.7 min | Failed · $1.14 · 12.2 min | Neither | GPT |
| `stet-commit-30506f8e` · 476 lines, 17 files | Failed · 71.2 min | Failed · $0.81 · 20.7 min | Neither | Even |
| `stet-commit-87814583` · 763 lines, 8 files | Failed · $0.38 · 61.5 min | Failed · $3.24 · 17.9 min | Neither | Even |

## Cost and time

- **Cost:** DeepSeek mean $0.43, median $0.25 from 17/18 tasks; GPT mean $0.94, median $0.75. DeepSeek cost less on 16 of 17 tasks; GPT on 1.
- **Agent time:** DeepSeek mean 48.4 minutes, median 46.6; GPT mean 10.5 minutes, median 8.9. GPT finished faster on all 18 tasks.

## Where each setup fits

### Small fixes

10 tasks, 2–84 changed lines in 1–2 files. DeepSeek solved 7/9 versus GPT’s 6/9 and cost 2.3× less; GPT was 6.5× faster.

| Metric | DeepSeek | GPT |
|---|---:|---:|
| Solved | 7/9 | 6/9 |
| Cost per task | $0.18 | $0.42 |
| Agent time per task | 34.5 min | 5.3 min |

### Large, cross-cutting changes

14 tasks, 113–1,420 changed lines in up to 19 files. Both solved 3/9. GPT was 4.0× faster; DeepSeek cost $0.70 per task from 8/9 tasks versus GPT at $1.46.

| Metric | DeepSeek | GPT |
|---|---:|---:|
| Solved | 3/9 | 3/9 |
| Cost per task | $0.70 from 8/9 tasks | $1.46 |
| Agent time per task | 62.3 min | 15.7 min |

## Side by side

| Metric | DeepSeek V4.1 Flash max | GPT 6.1 Sol medium |
|---|---|---|
| Harness | `deepseek-harness` | `codex` |
| Reasoning effort | max | medium |
| Rank | 8 of 18 | 4 of 18 |
| Stet score | 51.0, 95% range 42.1–60.1 | 54.7, 95% range 46.0–63.4 |
| Solved | 10/24: 8 passed, 2 judged equivalent | 11/24: 9 passed, 2 judged equivalent |
| Cost per task | $0.49 from 23/24 tasks | $1.10 |
| Agent time per task | 49.3 min | 11.8 min |

## Research links

- [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?** We found no clear quality difference: GPT won 4, DeepSeek won 1, 13 tied, and 6 were undecided. DeepSeek leads on cost at 2.2× cheaper; GPT leads on agent time at 4.2× faster.

**How were they tested?** Both agents worked on the same 24 merged changes from our private repository, from the state before each change. Repository tests and judges graded every patch against the merged change; no public benchmark scores were used.

## Test 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 from private `Stet-AI/stet`, replayed from the state immediately before each change.
- Data from November 19, 2026.
- Both setups came from leaderboard run `bmk_d4ba310fdf50db4ccb2092e3` of 18 setups.

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

### More comparisons

- [DeepSeek V4.1 Flash max vs Pareto 26.9 medium](https://www.stet.sh/leaderboard/compare/deepseek-v4-1-flash-max-vs-pareto-26-9-medium)
- [DeepSeek V4.1 Flash max vs DeepSeek V4.1 Flash (10/17) max](https://www.stet.sh/leaderboard/compare/deepseek-v4-1-flash-10-17-max-vs-deepseek-v4-1-flash-max)
- [DeepSeek V4.1 Flash max vs Grok 4.7 high](https://www.stet.sh/leaderboard/compare/deepseek-v4-1-flash-max-vs-grok-4-7-high)
- [DeepSeek V4.1 Flash max vs Haiku 5.5 max](https://www.stet.sh/leaderboard/compare/deepseek-v4-1-flash-max-vs-haiku-5-5-max)
- [DeepSeek V4.1 Flash max vs GPT 6 Sol medium](https://www.stet.sh/leaderboard/compare/deepseek-v4-1-flash-max-vs-gpt-6-sol-medium)
- [GPT 6.1 Sol medium vs Muse Wonder Full low](https://www.stet.sh/leaderboard/compare/gpt-6-1-sol-medium-vs-muse-wonder-full-low)
- [GPT 6.1 Sol medium vs GPT 6 Sol medium](https://www.stet.sh/leaderboard/compare/gpt-6-1-sol-medium-vs-gpt-6-sol-medium)
- [GPT 6.1 Sol medium vs Opus 5 medium](https://www.stet.sh/leaderboard/compare/gpt-6-1-sol-medium-vs-opus-5-medium)
- [GPT 6.1 Sol medium vs Grok 4.7 high](https://www.stet.sh/leaderboard/compare/gpt-6-1-sol-medium-vs-grok-4-7-high)
- [GPT 6.1 Sol medium vs SWE-2 Max high](https://www.stet.sh/leaderboard/compare/gpt-6-1-sol-medium-vs-swe-2-max-high)
- [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%2Fdeepseek-v4-1-flash-max-vs-gpt-6-1-sol-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%2Fdeepseek-v4-1-flash-max-vs-gpt-6-1-sol-medium). Opening a form does not execute a tool.