Identity
- Status
- retired
- Developer ID
deepseek-ai/DeepSeek-V4-Flash-Vision-Exp- Released
- Aug 21, 2026
- Version
- DeepSeek-V4-Flash-Vision-Exp
- License
- mit
DeepSeekV4 Flash Vision Exp
DeepSeek-V4-Flash-Vision-Exp is DeepSeek's first experimental multimodal model in the DeepSeek-V4 family, adding visual modules and continued training to DeepSeek-V4-Flash.
- Token context
- 1,049K tokens
- Inputs
- Text, Image
- Outputs
- Text
- Released
- Aug 21, 2026
Providers & Pricing
Estimate workload cost →| Provider | Input / 1M | Output / 1M | Cache read / 1M | Cache write / 1M | Observed | Evidence |
|---|---|---|---|---|---|---|
 DeepinfraAvailable | $0.44 | $1.32 | $0.014 | — | Oct 19, 2026 | Provider-reported |
DeepSeekAvailable | $0.22 | $0.66 | $0.007 | — | Sep 2, 2026 | Official fact |
Benchmarks
All benchmark results →| LiveBench2026-06-25-90f8d44da965 | 79.67 | 52,644 | 1,270 | deepseek-v4-flash-vision-expaverage per case | Recomputedwinner eligible | Third-party benchmark |
At a Glance
Model Facts
Capacity
- Token context
- 1,049K tokens
- Maximum output
- 393K tokens
- Total parameters
- 304.6B
- 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
- chat, generation, reasoning, structured outputs, tools
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-Flash-Vision-Exp, deepseek-v4-flash-vision-exp, DeepSeek-V4-Flash-Vision-Exp
Explicit Unknowns
- Knowledge cutoff
Primary Evidence
Sources and Observation Date
deepseek-ai/DeepSeek-V4-Flash-Vision-Exp developer model repositoryDeepSeek · observed Sep 2, 2026DeepSeek Models & PricingDeepSeek · observed Sep 10, 2026DeepSeek-V4-Flash-Vision-Exp Release: Multimodal API Now LiveDeepSeek · observed Sep 2, 2026deepseek-ai/DeepSeek-V4-Flash-Vision-Exp config.json at 6821d6ad3681DeepSeek · observed Sep 2, 2026deepseek-ai/DeepSeek-V4-Flash-Vision-Exp README.md at 6821d6ad3681DeepSeek · observed Sep 2, 2026deepseek-ai/DeepSeek-V4-Flash-Vision-Exp immutable Hub API metadataDeepSeek · observed Sep 2, 2026deepseek-ai/DeepSeek-V4-Flash-Vision-Exp tokenizer_config.json at 6821d6ad3681DeepSeek · observed Sep 2, 2026Deepinfra catalogDeepinfra · observed Oct 19, 2026Fireworks Ai catalogFireworks Ai · observed Oct 19, 2026
Comparable Peers
Related Model Comparisons
| A | Pair | B | Context |
|---|---|---|---|
DeepSeek V4 Flash Vision ExpDeepSeek | vs | Claude Fable 5.1Anthropic | cross-developer peersimage, text |
DeepSeek V4 Flash Vision ExpDeepSeek | vs |  Gemini 3.1 ProGoogle DeepMind | cross-developer peerstext |
DeepSeek V4 Flash Vision ExpDeepSeek | vs | DeepSeek V4.1 FlashDeepSeek | family variantsimage, text |
DeepSeek V4 Flash Vision ExpDeepSeek | vs |  Qwen3.8 MaxQwen | cross-developer peersimage, text |
DeepSeek V4 Flash Vision ExpDeepSeek | vs | cross-developer peersimage, text | |
DeepSeek V4 Flash Vision ExpDeepSeek | vs | DeepSeek V4 ProDeepSeek | family variantstext |
DeepSeek V4 Flash Vision ExpDeepSeek | vs | Qwen3.8 27BQwen | cross-developer peersimage, text |
DeepSeek V4 Flash Vision ExpDeepSeek | vs |  GLM 5.3 FlashZ.ai | cross-developer peersimage, text |
DeepSeek V4 Flash Vision ExpDeepSeek | vs | DeepSeek V4 FlashDeepSeek | family variantstext |
DeepSeek V4 Flash Vision ExpDeepSeek | vs | DeepSeek V4 Pro 0813DeepSeek | family variantstext |
DeepSeek V4 Flash Vision ExpDeepSeek | vs | Gemma 4 31BGoogle DeepMind | cross-developer peersimage, text |
DeepSeek V4 Flash Vision ExpDeepSeek | vs |  Muse Glimmer 30BMeta | cross-developer peersimage, text |
Questions
DeepSeek V4 Flash Vision Exp FAQs
What is DeepSeek V4 Flash Vision Exp?+
DeepSeek-V4-Flash-Vision-Exp is DeepSeek's first experimental multimodal model in the DeepSeek-V4 family, adding visual modules and continued training to DeepSeek-V4-Flash. It is developed by DeepSeek and its current sourced lifecycle status is retired.
What inputs and outputs does DeepSeek V4 Flash Vision Exp support?+
DeepSeek V4 Flash Vision Exp's current record lists Text and Image as input and Text as output.
How much context does DeepSeek V4 Flash Vision Exp support?+
The current record lists a 1,049K-token context window and a maximum output of 393K tokens.
Is DeepSeek V4 Flash Vision Exp available through an API?+
DeepSeek V4 Flash Vision Exp is marked as API-available. The directly sourced provider records currently include Deepinfra, DeepSeek, and Fireworks Ai.
How large is DeepSeek V4 Flash Vision Exp?+
The current record lists 304.6B total parameters and an unknown active parameter count. Its architecture is DeepseekV4ForCausalLM.
What capabilities and tasks does DeepSeek V4 Flash Vision Exp support?+
DeepSeek V4 Flash Vision Exp's recorded capabilities are chat, generation, reasoning, structured outputs, and tools. Its supported tasks are image, reasoning, and text.
How much does DeepSeek V4 Flash Vision Exp cost through an API?+
The lowest directly sourced prices currently attached to DeepSeek V4 Flash Vision Exp are $0.22 per million input tokens through DeepSeek and $0.66 per million output tokens through DeepSeek. Prices are provider-specific and should be checked against each cited observation date.
Can DeepSeek V4 Flash Vision Exp be self-hosted?+
DeepSeek V4 Flash Vision Exp is an open-weight model. Self-hosting is supported, and the recorded license is mit.
Does DeepSeek V4 Flash Vision Exp have related versions or aliases?+
No predecessor is recorded. No successor is recorded. Recorded aliases are deepseek-ai/DeepSeek-V4-Flash-Vision-Exp, deepseek-v4-flash-vision-exp, and DeepSeek-V4-Flash-Vision-Exp.
Which models can DeepSeek V4 Flash Vision Exp be compared with?+
The current compatible peer set includes Claude Fable 5.1, Gemini 3.1 Pro, DeepSeek V4.1 Flash, Qwen3.8 Max, and Llama 4 Maverick 17B 128E Instruct. Compatibility requires shared sourced modalities and tasks; it does not imply that one model is better. Browse model comparisons →