Identity
- Status
- active
- Developer ID
deepseek-ai/DeepSeek-V4.1-Flash- Released
- Sep 10, 2026
- Version
- DeepSeek-V4.1-Flash
- License
- mit
DeepSeekV4.1 Flash
DeepSeek-V4.1-Flash is DeepSeek's open-weight multimodal mixture-of-experts model with a 552B-parameter backbone, phase-dependent 8B/16B activation, one-million-token context, and continuously adjustable reasoning effort.
- Token context
- 1,049K tokens
- Inputs
- Text, Image
- Outputs
- Text
- Released
- Sep 10, 2026
Benchmark Market Position
Intelligence, Cost, and Efficiency
Not ranked in current cohort
Rank #5 of 44
Grok Build 0.1
GLM 5.3 Flash
MiniMax M3
Qwen3.8 Flash- DeepSeek V4.1 Flash
GPT-5.6 Luna
Gemini 3.5 Flash Lite
Kimi K2.7 Code- DeepSeek V4 Pro
- Gemini 3.6 Flash
- GLM 5.2
- Qwen3.7 Max
- GPT-5.4 Nano
- Qwen3.8 27B
- Gemini 3.7 Flash
Inkling
NVIDIA Nemotron 3 Ultra 550B A55B BF16- DeepSeek V4 Pro 0813
- Kimi K2.6
- Qwen3.8 Max
- GPT-5.6 Sol
- Gemini 3.5 Flash
- Gemini 3.8 Flash
- Gemini 3.1 Pro
- GPT-5.2
Claude Sonnet 5
- Kimi K3
- GPT-5.4 Mini
- GLM 5.3
- GPT-5.6 Terra
- GPT-5.4
- GPT-5.5
- GPT-6 Astra
Not ranked in current cohort
The orange outline marks DeepSeek V4.1 Flash. Each available panel preserves its real rank and value; missing benchmark or pricing inputs remain explicitly unranked.
Providers & Pricing
Estimate workload cost →| Provider | Input / 1M | Output / 1M | Cache read / 1M | Cache write / 1M | Observed | Evidence |
|---|---|---|---|---|---|---|
 DeepinfraAvailable | $0.20 | $0.60 | $0.006 | — | Sep 22, 2026 | Provider-reported |
DeepSeekAvailable | $0.15 | $0.60 | $0.003 | — | Sep 10, 2026 | Official fact |
Benchmarks
All benchmark results →| LiveBench2026-06-25 | 83.20 | 36,355 | 1,270 | deepseek-v4.1-flash-maxaverage per case | Recomputedwinner eligible | Third-party benchmark |
| LMArena Agent Arenaagent-2026-09-15-d25aabda0010 | 4.88 | — | 20,080 | Deepseek V4.1 Flash (Max); 95% CI [3.54852933, 6.21109128]; sessions 20080; observations 2295632; rank 12unknown | Original sourcewinner eligible | Third-party benchmark |
| ToneBench2026-09-11-10-task-4ef099199c9c | 88.03 | 18,199 | 50 | DeepSeek V4.1 Flash (max)average per case | Original sourcenot winner eligible | Third-party benchmark |
| ToneBench2026-09-11-10-task-4ef099199c9c | 87.13 | 10,011 | 50 | DeepSeek V4.1 Flash (default)average per case | Original sourcenot winner eligible | Third-party benchmark |
At a Glance
Model Facts
Capacity
- Token context
- 1,049K tokens
- Maximum output
- 393K tokens
- Total parameters
- 763.2B
- Active parameters
- Unknown
- Knowledge cutoff
- Not reported
Interface and Access
- Inputs
- Text, Image
- Outputs
- Text
- API available
- Yes
- Open weights
- Yes
- Self-hostable
- Yes
- Reasoning
- Yes
- Vision
- Yes
- Tool calling
- Yes
- Capabilities
- agents, chat, fim, generation, reasoning, responses, structured outputs, tools, vision
Technical Specifications
- Architecture design
- Causal Encoder-Decoder (20 encoder + 20 decoder layers)
- Backbone parameters
- 552,000,000,000 parameters
- Active parameters during prefill
- 8,000,000,000 parameters
- Active parameters during decode
- 16,000,000,000 parameters
- Transformer layers
- 40 layers
- Routed experts per MoE layer
- 384 experts
- Routed experts per token
- 6 experts
- Reasoning effort range
- 1–100
- Pre-training corpus
- 45,000,000,000,000 tokens
This page represents one developer model product. Dated API snapshots, serving endpoints, pricing, regions, quantizations, and service tiers attach as versioned aliases or provider details and never create duplicate public model records.
Version Lineage
PredecessorUnknown
SuccessorsUnknown
Aliasesdeepseek-ai/DeepSeek-V4.1-Flash, deepseek-flash, DeepSeek V4.1, deepseek-v4-flash, deepseek-v4-flash-vision-exp
Explicit Unknowns
- Knowledge cutoff
Primary Evidence
Sources and Observation Date
deepseek-ai/DeepSeek-V4.1-Flash developer model repositoryDeepSeek · observed Sep 10, 2026DeepSeek Models & PricingDeepSeek · observed Sep 10, 2026deepseek-ai/DeepSeek-V4.1-Flash config.json at dba1be0a40aaDeepSeek · observed Sep 10, 2026deepseek-ai/DeepSeek-V4.1-Flash README.md at dba1be0a40aaDeepSeek · observed Sep 10, 2026deepseek-ai/DeepSeek-V4.1-Flash LICENSE at dba1be0a40aaDeepSeek · observed Sep 10, 2026deepseek-ai/DeepSeek-V4.1-Flash immutable Hub API metadataDeepSeek · observed Sep 10, 2026deepseek-ai/DeepSeek-V4.1-Flash inference/README.md at dba1be0a40aaDeepSeek · observed Sep 10, 2026deepseek-ai/DeepSeek-V4.1-Flash tokenizer_config.json at dba1be0a40aaDeepSeek · observed Sep 10, 2026Deepinfra catalogDeepinfra · observed Sep 22, 2026Together Ai catalogTogether Ai · observed Sep 22, 2026
Comparable Peers
Related Model Comparisons
| A | Pair | B | Context |
|---|---|---|---|
DeepSeek V4.1 FlashDeepSeek | vs | DeepSeek V4 FlashDeepSeek | family variantstext |
DeepSeek V4.1 FlashDeepSeek | vs | GPT-6 AstraOpenAI | cross-developer peersimage, text |
DeepSeek V4.1 FlashDeepSeek | vs | Claude Fable 5.1Anthropic | cross-developer peersimage, text |
DeepSeek V4.1 FlashDeepSeek | vs | Gemini 3.1 ProGoogle DeepMind | cross-developer peerstext |
DeepSeek V4.1 FlashDeepSeek | vs | Grok 4.6xAI | cross-developer peersimage, text |
DeepSeek V4.1 FlashDeepSeek | vs | Qwen3.8 MaxQwen | cross-developer peersimage, text |
DeepSeek V4.1 FlashDeepSeek | vs | Kimi K3Moonshot AI | cross-developer peersimage, text |
DeepSeek V4.1 FlashDeepSeek | vs | MiniMax M3MiniMax | cross-developer peersimage, text |
DeepSeek V4.1 FlashDeepSeek | vs | GLM 5.3Z.ai | cross-developer peerstext |
DeepSeek V4.1 FlashDeepSeek | vs |  Hy4 previewTencent | cross-developer peerstext |
DeepSeek V4.1 FlashDeepSeek | vs | Seed 2.1 ProByteDance Seed | cross-developer peersimage, text |
DeepSeek V4.1 FlashDeepSeek | vs |  Mistral Large 3Mistral AI | cross-developer peersimage, text |
DeepSeek V4.1 FlashDeepSeek | vs | cross-developer peersimage, text | |
DeepSeek V4.1 FlashDeepSeek | vs |  Nova 2 LiteAmazon | cross-developer peersimage, text |
DeepSeek V4.1 FlashDeepSeek | vs | cross-developer peersimage, text | |
DeepSeek V4.1 FlashDeepSeek | vs |  ERNIE 5.1Baidu | cross-developer peerstext |
DeepSeek V4.1 FlashDeepSeek | vs | cross-developer peerstext | |
DeepSeek V4.1 FlashDeepSeek | vs | cross-developer peerstext | |
DeepSeek V4.1 FlashDeepSeek | vs |  Phi-4 ReasoningMicrosoft | cross-developer peerstext |
DeepSeek V4.1 FlashDeepSeek | vs | cross-developer peerstext | |
DeepSeek V4.1 FlashDeepSeek | vs |  Sonar ProPerplexity | cross-developer peerstext |
DeepSeek V4.1 FlashDeepSeek | vs | Claude Mythos 5.1Anthropic | cross-developer peersimage, text |
Questions
DeepSeek V4.1 Flash FAQs
What is DeepSeek V4.1 Flash?+
DeepSeek-V4.1-Flash is DeepSeek's open-weight multimodal mixture-of-experts model with a 552B-parameter backbone, phase-dependent 8B/16B activation, one-million-token context, and continuously adjustable reasoning effort. It is developed by DeepSeek and its current sourced lifecycle status is active.
What inputs and outputs does DeepSeek V4.1 Flash support?+
DeepSeek V4.1 Flash's current record lists Text and Image as input and Text as output.
How much context does DeepSeek V4.1 Flash support?+
The current record lists a 1,049K-token context window and a maximum output of 393K tokens.
What technical specifications are published for DeepSeek V4.1 Flash?+
DeepSeek V4.1 Flash's developer-published specifications include architecture design: Causal Encoder-Decoder (20 encoder + 20 decoder layers); backbone parameters: 552000000000 parameters; active parameters during prefill: 8000000000 parameters; active parameters during decode: 16000000000 parameters; transformer layers: 40 layers; routed experts per moe layer: 384 experts; routed experts per token: 6 experts; reasoning effort range: 1–100; pre-training corpus: 45000000000000 tokens.
How large is DeepSeek V4.1 Flash?+
The current record lists 763.2B total parameters and an unknown active parameter count. Its architecture is DeepseekV41ForCausalLM.
What capabilities and tasks does DeepSeek V4.1 Flash support?+
DeepSeek V4.1 Flash's recorded capabilities are agents, chat, fim, generation, reasoning, responses, structured outputs, tools, and vision. Its supported tasks are image, reasoning, and text.
Is DeepSeek V4.1 Flash available through an API?+
DeepSeek V4.1 Flash is marked as API-available. The directly sourced provider records currently include Deepinfra, DeepSeek, and Together Ai.
How much does DeepSeek V4.1 Flash cost through an API?+
The lowest directly sourced prices currently attached to DeepSeek V4.1 Flash are $0.15 per million input tokens through DeepSeek and $0.60 per million output tokens through Deepinfra. Prices are provider-specific and should be checked against each cited observation date.
Can DeepSeek V4.1 Flash be self-hosted?+
DeepSeek V4.1 Flash is an open-weight model. Self-hosting is supported, and the recorded license is mit.
Does DeepSeek V4.1 Flash have related versions or aliases?+
No predecessor is recorded. No successor is recorded. Recorded aliases are deepseek-ai/DeepSeek-V4.1-Flash, deepseek-flash, DeepSeek V4.1, deepseek-v4-flash, and deepseek-v4-flash-vision-exp.