RankingDeepSeek V4.1 Flash
DeepSeek V4.1 Flash
DeepSeek V4.1 Flash by DeepSeek is a DeepSeek V4.1 model, released 2026-09-10. This profile collects the published benchmark results and reported effort settings for this model. Configurations are compared on the effort page. Ranks are on the leaderboard.
6 published benchmark measures · 45 measured effort results · 1 benchmark families contribute across 1 task areas. 1 capability estimates available in the full profile. See all results ↓ · Compare published benchmarks →
Model evidence summary
This profile combines published settings. It is not a runnable configuration or a leaderboard rank. Compare measured configurations →
Performance profile
Capabilities
Filled points are supported; hollow points are preliminary. Lines stop at unknown capabilities. Exact values and sources follow below.
Scores estimate outcomes against a shared reference panel; they are not accuracy percentages. Sparse or disconnected evidence cannot qualify an overall profile. Open a capability to inspect its evidence.
Reported effort · Not specified
Settings reported in this model's published benchmark results, including results outside the aggregate. Effort names are provider-specific. These are not API defaults or equal compute budgets.
- Not specified: 6 observations
Mixed settings means multiple settings occur in the evidence. Best across efforts means the source selected its best reported result across settings; it does not mean Max. Unspecified settings stay unknown. This model-summary chart combines reported settings. The leaderboard keeps identified configurations separate and excludes unknown effort.
Inspect each result and its source ↓ · Download effort evidenceScore contributions and missing evidence
1 contributing families across 1 capabilities. Fixed reference panels do not change when the catalog expands.
Capability is fitted jointly across families. Capability estimates below describe different task areas; their weighted sum is not the Capability score.
Results without reviewed compatibility or a reference match remain in the raw evidence below. Coverage counts only contributing results.
Model information & shareable badge
- Lab
- DeepSeek
- Catalog status
- api-listed
- Availability
- Documented provider API; distinct September 10 checkpoint. Retired V4 Flash endpoint aliases now route here; historical checkpoints remain separate.
- Family
- DeepSeek V4.1
- Released
- 2026-09-10
- Context
- 1,000,000 tokens
- License
- —
- Model card
- https://api-docs.deepseek.com/updates/
- Default Capability family coverage
/badge/deepseek-v4.1-flash.svg
DeepSeek V4.1 Flash measured results by effort setting
Each row is one published result for one effort setting. Evidence collected 2026-11-22. Compare the effort settings →
| Benchmark | Effort | Score | Source |
|---|---|---|---|
| AA-AnalystAgent | Max | 35.0% | Artificial Analysis |
| AA-Briefcase Elo | Max | 1415 | Artificial Analysis |
| AA-Briefcase Elo | Non-reasoning | 1118 | Artificial Analysis |
| AA-LCR v1.1 | Max | 84.0% | Artificial Analysis |
| AA-LCR v1.1 | Non-reasoning | 55.3% | Artificial Analysis |
| ARC-AGI-2 | High | 67.5% | ARC Prize |
| ARC-AGI-2 | Low | 60.6% | ARC Prize |
| ARC-AGI-2 | Max | 72.9% | ARC Prize |
| ARC-AGI-2 | None | 1.3% | ARC Prize |
| ARC-AGI-3 · provider-adapter | High | 3.2% | ARC Prize |
| ARC-AGI-3 · provider-adapter | Low | 3.3% | ARC Prize |
| ARC-AGI-3 · provider-adapter | Max | 2.7% | ARC Prize |
| ARC-AGI-3 · provider-adapter | None | 0.5% | ARC Prize |
| ARC-AGI-3 · standard | High | 0.5% | ARC Prize |
| ARC-AGI-3 · standard | Low | 0.3% | ARC Prize |
| ARC-AGI-3 · standard | Max | 0.9% | ARC Prize |
| ARC-AGI-3 · standard | None | 0.3% | ARC Prize |
| Artificial Analysis Cyber Index | Max | 40.8 | Artificial Analysis |
| AutomationBench-AA | Max | 68.9% | Artificial Analysis |
| AutomationBench-AA | Non-reasoning | 47.8% | Artificial Analysis |
| CritPt | Max | 14.3% | Artificial Analysis |
| CritPt | Non-reasoning | 0.3% | Artificial Analysis |
| FrontierCode 1.1 · extended-devin-cli | High | 56.6% | Cognition |
| FrontierCode 1.1 · main-devin-cli | High | 43.9% | Cognition |
| FrontierCS 2 | Unspecified | 14.6% | FrontierCS |
| GDP.pdf: All-pass Rate | Max | 12.8% | Artificial Analysis |
| GDP.pdf: All-pass Rate | Non-reasoning | 3.8% | Artificial Analysis |
| GDPval-AA v2.1 Leaderboard | Max | 1600 | Artificial Analysis |
| GDPval-AA v2.1 Leaderboard | Non-reasoning | 1347 | Artificial Analysis |
| Harvey LAB-AA v1.1: Hallucination-Gated All-Pass Rate | Max | 1.7% | Artificial Analysis |
| Humanity's Last Exam | Max | 39.2% | Artificial Analysis |
| Humanity's Last Exam | Non-reasoning | 10.8% | Artificial Analysis |
| ITBench-AA | Max | 46.9% | Artificial Analysis |
| LMArena Agent Arena | Max | 4.0 | LMArena |
| MaxxBench | Max | 55.6% | unLAP |
| MLCR-AA | Max | 23.9% | Artificial Analysis |
| MMMU-Pro | Max | 77.0% | Artificial Analysis |
| MMMU-Pro | Non-reasoning | 65.6% | Artificial Analysis |
| Multi-step SWE | Max | 6.5% | Klavis AI |
| SciCode | Max | 51.9% | Artificial Analysis |
| SciCode | Non-reasoning | 35.5% | Artificial Analysis |
| Terminal-Bench v4.0 | Max | 26.8% | Artificial Analysis |
| Terminal-Bench v4.0 | Non-reasoning | 5.6% | Artificial Analysis |
| Terminal-Bench-Science 0.1 | Max | 9.0% | Artificial Analysis |
| Terminal-Bench-Science 0.1 | Non-reasoning | 0.0% | Artificial Analysis |
Benchmark scores & sources
Original results, evaluation harnesses, and evidence behind this model.
| Benchmark | Bucket | Score | Harness | Evidence | Source-recorded date |
|---|---|---|---|---|---|
| ARC-AGI-2 | Hard reasoning | 72.9% | ARC Prize verifiedcontributes to capability | official board | 2026-09-10 |
| Artificial Analysis Cyber Index | Supporting evidence | 40.81131888823 | Artificial Analysis Stirrup · Cyber Index v1.0 · maxCyber Index source settingsArtificial Analysis Stirrup, v1.0. Display only. This composite measures defensive source-code tasks. Successful capability scores depend on safety refusals, and refusal rates are reported separately. It does not enter weighted Overall. Official Cyber Index board | official board | 2026-11-18 |
| Bug Hunt Bench | Supporting evidence | 20.67% | Bug Hunt Bench · max | official board | 2026-09-15 |
| FrontierCS 2 | Supporting evidence | 12.5% | DeepSeek Harness | official board | 2026-11-20 |
| LMArena Agent Arena | Supporting evidence | 2.74 | LMArena Agent Arena · max | official board | 2026-10-16 |
| Multi-step SWE | Supporting evidence | 6.5% | Claude Code | official board | 2026-11-06 |
| DeepSWE v1.1 | Agentic | — | — | — | — |
| GDPval-AA | Agentic | — | — | — | — |
| GPQA Diamond | Hard reasoning | — | — | — | — |
| Humanity's Last Exam | Hard reasoning | — | — | — | — |
| LiveCodeBench | Coding | — | — | — | — |
| LMArena Text Arena | Human pref | — | — | — | — |
| MMLU-Pro | Knowledge | — | — | — | — |
| OSWorld-Verified | Agentic | — | — | — | — |
| SWE-bench Pro | Agentic | — | — | — | — |
| SWE-bench Verified | Agentic | — | — | — | — |
| Terminal-Bench 2.1 | Agentic | — | — | — | — |