# DeepSeek V4.1 Flash - BenchAudit True Index Profile

**Integrity Profile: Clean Generalization**

> The Verdict: Performs Better on Fresh Private Evals Than Legacy Tests. DeepSeek V4.1 Flash (evaluated at max effort) outperforms its legacy benchmark standing (39.0p) when moved to private, leak-resistant evaluations (48.3p), posting a -6.9p Low-Tier Boost across 3 matched domains, particularly in Coding & SWE (-18.4p boost) and Agentic & Enterprise (-13.5p boost)—a strong indicator of genuine generalization.

- **Canonical URL:** https://benchaudit.org/models/deepseek-v4-1-flash__max/
- **True Index Rank:** #26 of 44
- **Creator:** DeepSeek
- **Weights:** Open Weights
- **Clean Evals Avg (Tiers S & A):** 48.3p (7 evals)
- **Legacy Benchmarks Avg (Tiers C & D):** 39.0p (6 evals)
- **Low-Tier Boost (Matched Delta):** -6.9p
- **Updated:** 2026-11-22

## True Index: Domain Rankings

| Capability Domain | Sub-Rank | Clean Evals (S/A) Avg | Top-Tier Evals |
| --- | ---: | ---: | ---: |
| Agentic & Enterprise | #33 | 41.4p | 4 |
| Coding & SWE | #20 | 59.2p | 1 |
| Math & Scientific Reasoning | #28 | 44.1p | 2 |

## Head-to-Head: Clean Evals vs. Legacy Benchmarks

| Capability Domain | Clean Evals (Tiers S & A) | Legacy Benchmarks (Tiers C & D) | Public Lift (Boost) |
| --- | ---: | ---: | ---: |
| Agentic & Enterprise | 4 (41.4p) | 2 (27.9p) | -13.5p |
| Coding & SWE | 1 (59.2p) | 1 (40.8p) | -18.4p |
| Math & Scientific Reasoning | 2 (44.1p) | 2 (68.9p) | +24.8p |

## Audited Benchmark Performance of DeepSeek V4.1 Flash

| Benchmark | Benchmark Integrity Status | Domain | Rank | Percentile |
| --- | --- | --- | ---: | ---: |
| [HealthBench](https://benchaudit.org/benchmarks/healthbench/) | Public / Aging | Math & Scientific Reasoning | #2/16 | 98.1p |
| [AutomationBench-AA](https://benchaudit.org/benchmarks/automationbench-aa/) | Private / Fresh | Agentic & Enterprise | #14/50 | 73.5p |
| [MLCR-AA](https://benchaudit.org/benchmarks/mlcr-aa/) | Private / Fresh | Math & Scientific Reasoning | #17/46 | 67.1p |
| [Code2Video-AA](https://benchaudit.org/benchmarks/code2video-aa/) | Private / Fresh | Coding & SWE | #9/23 | 59.2p |
| [Harvey LAB-AA v1.1](https://benchaudit.org/benchmarks/harvey-lab-aa-v1-1/) | Private / Fresh | Agentic & Enterprise | #17/21 | 44.8p |
| [Terminal-Bench 4.0](https://benchaudit.org/benchmarks/terminal-bench-4-0/) | Public / Aging | Coding & SWE | #30/50 | 40.8p |
| [Terminal-Bench-Science 0.1](https://benchaudit.org/benchmarks/terminal-bench-science-0-1/) | Public / Aging | Math & Scientific Reasoning | #27/41 | 39.7p |
| [GDPval-AA v2.1](https://benchaudit.org/benchmarks/gdpval-aa-v2-1/) | Public / Aging | Agentic & Enterprise | #26/43 | 37.4p |
| [HLE-Diamond](https://benchaudit.org/benchmarks/hle-diamond/) | Semi-Exposed | General Knowledge & QA | #36/49 | 27.1p |
| [AA-Briefcase v1.1](https://benchaudit.org/benchmarks/aa-briefcase-v1-1/) | Private / Fresh | Agentic & Enterprise | #37/50 | 26.5p |
| [CritPt v2](https://benchaudit.org/benchmarks/critpt-v2/) | Private / Fresh | Math & Scientific Reasoning | #30/40 | 21.2p |
| [AA-AnalystAgent](https://benchaudit.org/benchmarks/aa-analystagent/) | Private / Fresh | Agentic & Enterprise | #22/24 | 20.9p |
| [GDP.pdf](https://benchaudit.org/benchmarks/gdp-pdf/) | Public / Aging | Agentic & Enterprise | #41/50 | 18.4p |
| [MMMU-Pro](https://benchaudit.org/benchmarks/mmmu-pro/) | Public / Aging | General Knowledge & QA | #18/22 | 18.2p |
