# DeepSeek V4.1 Flash max vs Grok 4.7 high

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-grok-4-7-high
Source observed: 2026-11-19T11:08:43.198Z
Last checked: 2026-11-19T11:08:43.198+00:00
Indexed: 2026-11-19T11:14:49.032+00:00

---

# DeepSeek V4.1 Flash max vs Grok 4.7 high

We compare DeepSeek V4.1 Flash max and Grok 4.7 high as coding-agent setups on real repository work. Grok is faster; DeepSeek is cheaper; the headline quality result is even. DeepSeek used `deepseek-harness` with max reasoning, while Grok used `cursor` with high reasoning.

## Headline results

- **Quality:** DeepSeek won 3, Grok won 4, 11 were tied, and 6 were undecided.
- **Speed:** Grok was 2.6× faster: 18.9 versus 49.3 minutes per task.
- **Cost:** DeepSeek was 13× cheaper: $0.49 versus $6.28 per task. DeepSeek’s figure is from 23/24 tasks; Grok’s from 22/24.

## Test design and scores

We ran both setups on the same 24 merged changes from private `Stet-AI/stet`, each from the repository state before the change. Repository tests and judges compared each patch with the merged change. Data is from November 19, 2026.

| Metric | DeepSeek V4.1 Flash max | Grok 4.7 high |
|---|---:|---:|
| Harness | `deepseek-harness` | `cursor` |
| Reasoning effort | max | high |
| Rank | 8 of 18 | 6 of 18 |
| Stet score | 51.0; 95% range 42.1–60.1 | 52.7; 95% range 45.4–59.9 |
| Solved | 10/24: 8 passed, 2 equivalent | 11/24: 9 passed, 2 equivalent |
| Cost / task | $0.49, from 23/24 tasks | $6.28, from 22/24 tasks |
| Agent time / task | 49.3 min | 18.9 min |

The score ranges overlap. Stet score is from 0 to 100, with 50 even with the field.

### Other quality measures

| Metric | DeepSeek | Grok |
|---|---:|---:|
| Code review pass rate | 33% — 24 tasks | 39% — 23 tasks |
| Checklist score | 0.83 — 24 tasks | 0.84 — 24 tasks |
| Equivalence rate | 43% — 23 tasks | 39% — 23 tasks |

These measures are outside the headline score and have no confidence interval.

## Task breakdown

DeepSeek solved 3 tasks Grok did not; Grok solved 2 DeepSeek did not. Both solved 7; neither solved 6. Among the 7 both solved, judges preferred Grok’s patch on 2 and called 5 even.

### Only DeepSeek solved

- 4 lines, 1 file — `stet-43493c52`; DeepSeek patch preferred
- 5 lines, 2 files — `stet-15439c21`; DeepSeek patch preferred
- 640 lines, 8 files — `stet-commit-342d0ac0`; DeepSeek patch preferred

### Only Grok solved

- 2 lines, 1 file — `stet-2450ca2d`; Grok patch preferred
- 113 lines, 4 files — `stet-commit-f0caded9`; Grok patch preferred

### Both solved

- 9 lines, 1 file — `stet-89dfbc27`
- 10 lines, 1 file — `stet-232b50ba`
- 23 lines, 2 files — `stet-5276aa3f`
- 29 lines, 1 file — `stet-b28f4fe7`
- 84 lines, 1 file — `stet-66762b4c`
- 218 lines, 2 files — `stet-commit-5d766dde`; Grok patch preferred
- 1,312 lines, 9 files — `stet-commit-f8ae1177`; Grok patch preferred

### Neither solved

- 23 lines, 1 file — `stet-6f84e978`
- 149 lines, 2 files — `stet-commit-e928166f`
- 476 lines, 17 files — `stet-commit-30506f8e`
- 552 lines, 11 files — `stet-commit-60a0e6c2`
- 763 lines, 8 files — `stet-commit-87814583`
- 1,420 lines, 19 files — `stet-commit-2826765e`

Six of 24 tasks are excluded from the detailed comparison because one setup has no scored result. A task counts as solved when tests pass or a judge finds the fix equivalent; Stet score also weighs code quality.

## Cost and time on scored tasks

On 18 scored tasks, DeepSeek averaged $0.43, median $0.25, from 17/18 tasks; Grok averaged $5.00, median $2.53, from 17/18. DeepSeek cost less on all 16 tasks with comparable cost data. DeepSeek averaged 48.4 minutes, median 46.6; Grok averaged 14.6, median 12.4. Grok finished faster on all 18 tasks.

### Small fixes

DeepSeek solved 7/9 versus Grok 6/9 at 9.4× lower cost; Grok was 4.3× faster.

|  | DeepSeek | Grok |
|---|---:|---:|
| Solved | 7/9 | 6/9 |
| Cost / task | $0.18 | $1.74 |
| Agent time / task | 34.5 min | 8.0 min |

### Large, cross-cutting changes

Both solved 3/9. Grok was 2.9× faster.

|  | DeepSeek | Grok |
|---|---:|---:|
| Solved | 3/9 | 3/9 |
| Cost / task | $0.70, from 8/9 tasks | $8.68, from 8/9 tasks |
| Agent time / task | 62.3 min | 21.2 min |

## Questions and next step

We found no clear quality difference: Grok won 4, DeepSeek won 3, 11 tied, and 6 were undecided. DeepSeek leads on cost; Grok leads on time. [Get started](https://www.stet.sh/private) to replay your merged changes and receive comparison reports through your agent.

[Leaderboard](https://www.stet.sh/leaderboard) · [How Stet measures](https://www.stet.sh/methodology) · [Cost per successful task](https://www.stet.sh/blog/cost-per-successful-task) · [Benchmark mined from our repository](https://www.stet.sh/blog/benchmark-agreed-with-frontier-code) · [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-grok-4-7-high&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-grok-4-7-high). Opening a form does not execute a tool.