deepseek
DeepSeek: DeepSeek V4 Flash Vision Exp
deepseek/deepseek-v4-flash-vision-exp
Our take
AI-assisted analysis based on the sources below. Suitability is an editorial assessment; test it on your own workload.
DeepSeek V4 Flash Vision Exp is an experimental multimodal variant in DeepSeek's V4 Flash family that accepts text and image inputs and returns text. OpenRouter lists a roughly one-million-token context, configurable reasoning effort, tool calling, and structured-output controls. DeepSeek says visual modules and continued training add image understanding, and its reported evaluations show substantial gains over DeepSeek-V4-Flash-0731 on multimodal agent tasks with comparable text-agent performance. The evidence is limited to catalog data and DeepSeek's own documentation, so the reported results need validation on real workloads.[1][2]
Strengths
- Accepts image and text inputs and emits text; DeepSeek describes it as its first experimental multimodal model in the V4 family, built with visual modules and continued training.[1][2]
- OpenRouter lists a context window of roughly one million tokens, plus tool use, structured outputs, and selectable reasoning effort including a maximum setting.[1]
- DeepSeek-reported benchmark results indicate stronger multimodal agent performance than DeepSeek-V4-Flash-0731 while preserving comparable text-agent performance; these are vendor measurements, not independent results.[2]
Limitations
- OpenRouter lists text as the only output modality; this catalog entry does not cover image generation or other output types.[1]
- The model is explicitly labeled experimental; the pack includes catalog data and DeepSeek's own documentation, but no independent evaluation or production-readiness evidence. Validate accuracy, safety, cost, and latency on the intended workload.[1][2]
- The knowledge cutoff is not provided in the catalog data.[1]
Where to use it
- Consider it for prototypes or workflows that need question answering or visual analysis from image inputs where a text-only response is sufficient.[1][2]
- Evaluate it for coding assistants or agents that combine visual inputs with tool calls and structured responses.[1][2]
- Try it for long-context multimodal applications that can benefit from the cataloged context window, after confirming deployment support and cost.[1][2]
Sources
- [1] OpenRouter catalog · DeepSeek: DeepSeek V4 Flash Vision ExpSources checked: Nov 13, 2026
- [2] Model documentation · DeepSeek: DeepSeek V4 Flash Vision ExpSources checked: Nov 13, 2026
About this model
Description below is the original text supplied by OpenRouter, shown in the source language.
DeepSeek V4 Flash Vision Exp is an experimental vision-enabled version of [DeepSeek V4 Flash 0731](https://openrouter.ai/deepseek/deepseek-v4-flash-0731) from DeepSeek, adding image understanding while matching the base model on text capabilities including agents,...
Specifications
- Model ID
- deepseek/deepseek-v4-flash-vision-exp
- Canonical ID
- deepseek/deepseek-v4-flash-vision-exp-20260821
- Creator
- deepseek
- Added to OpenRouter
- 21.08.2026
- Context length
- 1,048,576
- Modality
- text+image->text
- Input modalities
- text, image
- Output modalities
- text
- Token pricing
- USD per 1M tokens: in $0.440 / out $1.32
- Knowledge cutoff
- —
- Expiration date
- —
Tasks and capabilities
A grade appears only where the source supplies a matching measurement for that task, the model is active and the cohort holds at least 10 comparable models. The capability column lists facts the source declares — supported formats, reasoning and tool parameters — which are not measured quality.
| Task | Grade | Source-declared capability |
|---|---|---|
| General quality | No measurement | reasoning: optional · default high · efforts: max, high, low, none · parameters: include_reasoning, reasoning, reasoning_effort |
| Coding | AA — next 20% of the cohort | — |
| Agents | AA — next 20% of the cohort | parameters: structured_outputs, tool_choice, tools |
| Writing | No measurement | output: text |
| Reasoning | No measurement | reasoning: optional · default high · efforts: max, high, low, none · parameters: include_reasoning, reasoning, reasoning_effort |
| Document analysis | No measurement | — |
| Vision | No measurement | input: image |
| Image generation | No measurement | — |
| Audio | No measurement | — |
| Video | No measurement | — |
| Embeddings | No measurement | — |
Evaluations
| Source | Benchmark | Value | Grade | Data retrieved |
|---|---|---|---|---|
| Artificial Analysis | Coding index | 65 | AA — next 20% of the cohort | 13.11.2026 |
| Artificial Analysis | Agentic index | 47.5 | AA — next 20% of the cohort | 13.11.2026 |
How grades work
Each letter grade compares a model only against other models scored by the same source and metric in the current snapshot. S means the top 10%, A the next 20%, B the next 30%, C the rest. Models without a published measurement are shown as not rated — a supported format alone never implies measured quality.
- S — S — top 10% of the cohort
- A — A — next 20% of the cohort
- B — B — next 30% of the cohort
- C — C — below the B band
Providers
Optional live provider statistics from OpenRouter. Missing values are unknown, not zero.
| Provider | Context length | Price in /1M | Price out /1M | Latency p50 (30m, s) | Throughput p50 (30m, tok/s) | Uptime (30m) |
|---|---|---|---|---|---|---|
| DeepInfrafp8 | 1,048,576 | $0.216 | $0.647 | — | — | 100% |
| GMICloudfp8 | 1,048,575 | $0.440 | $1.32 | — | — | 100% |
Provider data retrieved: 13.11.2026
Related articles
No published articles identify this model yet.