deepseek-ai/DeepSeek-V4.1-Flash
image-text-to-text model by deepseek-ai
From the publisher
Model card & documentation
Source previewRead the publisher’s intended use, setup instructions, evaluations and limitations. The original model card is the source of truth.
Technical Report 👁️ We introduce DeepSeek-V4.1-Flash, a multimodal Mixture-of-Experts (MoE) model with 552B backbone parameters and support for contexts of up to one million tokens. The model natively processes images and text, and generates text autoregressively. Architecture. DeepSeek-V4.1-Flash adopts a Causal Encoder-Decoder (CED) architecture: a 40-layer Transformer organized as a 20-layer causal encoder followed by a 20-layer decoder. With CED, the decoder's global KV cache is projected from the final encoder hidden states rather than derived from each decoder layer's own hidden states. This allows the model to activate only 8B parameters per token during prefill and 16B during decode, substantially improving cost efficiency for input-heavy agentic workloads. SWA Bounded Replay reconstructs missing SWA KV states by replaying only the most recent nwin tokens, avoiding the need to persist SWA KV to SSD and reducing the persistent KV cache footprint to roughly 1/8 of that of DeepSeek-V4-Flash. Compressed Sparse Attention…
Read the full model card ↗ · Preview checked 2026-10-31T06:41:09.716Z
Inside the original model card — Document outline
- DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression
- Introduction
- Evaluation Results
- Base Model
- Instruct Model
- Comparison with frontier models (Max reasoning effort)
- Performance across agent scaffolds (DeepSWE v1.1 and Terminal-Bench 2.1, Max reasoning effort)
- Prompt Encoding
- Minimal Inference
- Reproducing DeepSWE Benchmark Results
- License
- Citation
Headings are captured from the source. Links open the publisher’s document, not a locally hosted copy.
Links from the model card
References supplied by the publisher, not independently verified endorsements. Check the destination before downloading files or entering credentials.
Documentation belongs to its respective authors. Reported project/model license: mit. A listing is not a grant of reuse or training rights. Confirm the document’s own terms at the source.
Model overview
deepseek-ai/DeepSeek-V4.1-Flash is a image-text-to-text model repository published by deepseek-ai on Hugging Face. The source reports the transformers library.
This page summarizes Hub metadata. For intended use, training data, evaluation results and limitations, consult the original model card.
Model facts
- Task
- image-text-to-text
- Library
- transformers
- Recent downloads (30 days)
- 2,071,064
- Cumulative likes
- 4,827
- Hugging Face trending score
- 78
- Reported safetensors parameters
- 763,205,315,794
- Architecture
- DeepseekV41ForCausalLM
- License
- mit
- Access
- Not gated by Hugging Face
- Created
- 2026-09-10T02:17:58.000Z
- Last modified
- 2026-10-16T09:27:28.000Z
Compatibility and lineage
Language coverage was not reported.
Base-model lineage was not included.
Parameter count is not a RAM/VRAM requirement. Check precision, quantization, context length and runtime compatibility in the model card; no hardware or API-cost claim is inferred here.
Use and evaluate
Check license terms and access requirements first. Review the model card’s documented loading instructions, evaluations and safety limitations. Benchmark scores are not imported by this directory, and popularity is not an accuracy ranking.
No model weights are downloaded or executed by AltAPIs. Never enable remote model code without reviewing it.
Source and freshness
Source: Hugging Face Hub. Metadata observed 2026-11-01T06:51:30.840Z. Daily imports are snapshots, not real-time monitoring.
Popularity and source listings do not establish security, suitability, licensing rights or benchmark performance.