Contamination MapText / LLM / Model
DeepSeek-V4.1-Flash
- Organisation
- DeepSeek-AI
- Family
- DeepSeek · version V4.1-Flash
- Released
- 2026-09-10 (DeepSeek API Docs news entry 'DeepSeek-V4.1-Flash Release 2026/09/10', https://api-docs.deepseek.com/news/news1226)
- Checkpoints
- deepseek-ai/DeepSeek-V4.1-Flash
- Weights
- Open · MIT
- Links
- PaperModel card
Training data disclosure: Vague
Level determined for 0% of benchmark cells.
- TestTest data seen0
- SeenSeen, split unclear0
- TrainTraining data seen0
- Related data seen0
- NoNot seen0
- ?Unknown15
“DeepSeek-V4.1-Flash is trained from scratch on a multimodal corpus comprising **45T tokens**”
Other declared training data
-
Common Crawl (multimodal web crawl: image-text pairs with alt text, interleaved webpages and PDFs) (broad web data: may contain benchmark test sets) · split: unknown
Evidence
“we therefore re-bootstrapped it from Common Crawl to improve its coverage of multimodal sources.”
huggingface.co · paper · Section 4.1 (Multimodal Data Curation) · retrieved 2026-10-15 “For interleaved data, we build this subset predominantly from webpages and PDFs.”
huggingface.co · paper · Section 4.1 · retrieved 2026-10-15 -
text corpus incl. recent code from open-source repositories, commits and libraries (broad web data: may contain benchmark test sets) · split: unknown
Evidence
“the new corpus incorporates more recent code from newly released open-source repositories, commits, libraries, and emerging frameworks”
huggingface.co · paper · Section 4.1 (Text Data Curation) · retrieved 2026-10-15 -
domain-specific vision datasets, image-code pairs and computer-use trajectories (not yet matched to a dataset) · split: unknown
Evidence
“We also collect extensive image-code pairs and computer-use trajectories to improve multimodal agentic understanding.”
huggingface.co · paper · Section 4.1 · retrieved 2026-10-15
Benchmark cells
Classic LLM benchmarks
Unknown MMLU (Measuring Massive Multitask Language Understanding) · training data not disclosed in enough detail
Unknown GSM8K (Grade School Math 8K) · training data not disclosed in enough detail
Unknown MATH (Mathematics Aptitude Test of Heuristics) · training data not disclosed in enough detail
Unknown HumanEval · training data not disclosed in enough detail
Unknown ARC-Challenge (AI2 Reasoning Challenge, Challenge Set) · training data not disclosed in enough detail
Unknown HellaSwag · training data not disclosed in enough detail
Unknown TruthfulQA · training data not disclosed in enough detail
Unknown WinoGrande · training data not disclosed in enough detail
Recent LLM benchmarks
Unknown GPQA Diamond · training data not disclosed in enough detail
Unknown MMLU-Pro · training data not disclosed in enough detail
Unknown LiveCodeBench (code generation, release_v6) · training data not disclosed in enough detail
Unknown SWE-bench Verified · training data not disclosed in enough detail
Unknown AIME 2024 · training data not disclosed in enough detail
Unknown AIME 2025 · training data not disclosed in enough detail
Unknown MATH-500 · training data not disclosed in enough detail
Notes
Native multimodal model trained from scratch; text-only to multimodal token ratio 7:1. No cutoff or decontamination statement found.
Sources
- DeepSeek-V4.1-Flash model card · model card · retrieved 2026-10-15
- DeepSeek-V4.1-Flash technical report · paper · retrieved 2026-10-15