DeepSeek V4 Flash Vision Exp

DeepSeek V4 Flash Vision Exp is an experimental vision-enabled version of DeepSeek V4 Flash 0731 from DeepSeek, adding image understanding while matching the base model on text capabilities including agents, reasoning, and world knowledge. It is a sparse mixture-of-experts model with 13B active parameters out of 284B total.

It is suited for document and chart understanding, visual question answering, and multimodal agent workflows that interleave text and images.

deepseek/deepseek-v4-flash-vision-exp

Organization: DeepSeek
Family: deepseek-flash
Providers: 15
Context: 1,000,000
Output limit: 384,000
Knowledge: —
Release: 2026-08-21
Updated: 2026-09-01
Weights: Open
Input: text, image/jpeg, image/png, image/gif, image/webp
Output: text
Capabilities: —

Providers

Compare every cataloged offer for this model, including its provider, route, token limits, price, and capabilities.

18 offers · USD per million tokens

Provider	Model ID	Context	Output	Input / output · 1M	Reasoning	Tools	Structured	Details
AIHubMix	deepseek-v4-flash-vision-exp	1,000,000	384,000	— / —	—	—	—	View
AI/ML API	deepseek/deepseek-v4-flash-vision-exp	1,048,576	384,000	— / —	—	—	—	View
AMD	DeepSeek-V4-Flash-Vision-Exp	1,000,000	384,000	— / —	—	—	—	View
CommandCode	deepseek/deepseek-v4-flash-vision-exp	1,000,000	384,000	— / —	—	—	—	View
CrossModel	deepseek/deepseek-v4-flash-vision-exp	1,000,000	384,000	— / —	—	—	—	View
Deep Infra	deepseek-ai/DeepSeek-V4-Flash-Vision-Exp	1,048,576	384,000	— / —	—	—	—	View
Eden AI	deepseek-v4-flash-vision-exp	1,000,000	384,000	— / —	—	—	—	View
Eden AI	deepseek/deepseek-v4-flash-vision-exp	1,000,000	384,000	— / —	—	—	—	View
Hugging Face	deepseek-ai/DeepSeek-V4-Flash-Vision-Exp	1,048,576	384,000	— / —	—	—	—	View
Kilo Gateway	deepseek/deepseek-v4-flash-vision-exp	1,048,576	262,144	— / —	—	—	—	View
NanoGPT	deepseek/deepseek-v4-flash-vision-exp	1,048,576	384,000	— / —	—	—	—	View
OpenCode	deepseek-v4-flash-vision-exp	1,000,000	384,000	— / —	—	—	—	View
DeepInfravia OpenRouter	deepseek/deepseek-v4-flash-vision-exp fp8	1,048,576	262,144	$0.2156 / $0.6468	Yes	Yes	Yes	View
GMICloudvia OpenRouter	deepseek/deepseek-v4-flash-vision-exp fp8	1,048,575	943,717	$0.44 / $1.32	Yes	Yes	No	View
SiliconFlowvia OpenRouter	deepseek/deepseek-v4-flash-vision-exp fp8	1,048,576	393,216	$0.44 / $1.32	Yes	Yes	No	View
OrcaRouter	deepseek/deepseek-v4-flash-vision-exp	1,000,000	384,000	— / —	—	—	—	View
SiliconFlow	deepseek-ai/DeepSeek-V4-Flash-Vision-Exp	1,000,000	384,000	— / —	—	—	—	View
Vercel AI Gateway	deepseek/deepseek-v4-flash-vision-exp	1,048,576	1,048,576	— / —	—	—	—	View

Prices and limits apply to each offer. A dash means the value is not provided. OpenRouter offers show the hosting provider and routing channel separately.

Pricing

Compare token rates and cache charges across available routes and hosting configurations.

USD per million tokens

Input from$0.2156Per 1M input tokens

Output from$0.6468Per 1M output tokens

Available offers18Across 15 providers

Lowest listed input and output prices across the offers above; they may belong to different providers. These are base rates, before conditional pricing, discounts or additional fees.

Endpoint price comparisonLatest observed snapshot · 3 lowest-priced configurations

InputOutput

DeepInfradeepinfra/fp8

$0.2156$0.6468

GMICloudgmicloud/fp8

$0.44$1.32

SiliconFlowsiliconflow/fp8

$0.44$1.32

deepseek/deepseek-v4-flash-vision-expObserved Oct 10, 2026, 11:31 AM UTC

Input: $0.2156
Output: $0.6468
Cache read: $0.00686

Hosting prices through OpenRouter All endpoints and cache rates

Endpoint	Input / 1M	Output / 1M	Cache read / 1M	Cache write / 1M
DeepInfradeepinfra/fp8	$0.2156	$0.6468	$0.00686	—
GMICloudgmicloud/fp8	$0.44	$1.32	$0.014	—
SiliconFlowsiliconflow/fp8	$0.44	$1.32	$0.028	—

Each row is a hosting configuration. Open an offer in Providers for its full pricing conditions and additional fees. Endpoint observations may differ from the routing catalog's base price.

Performance

Review reported latency and output-throughput percentiles from the latest observed rolling window.

Latest 30-minute observation

Time to first token · milliseconds

Latency distributionRolling 30-minute snapshot · 3 endpoints

P50P75P90P99

DeepInfradeepinfra/fp8

1,358.5–9,657.51 ms

GMICloudgmicloud/fp8

2,476.5–6,127.76 ms

SiliconFlowsiliconflow/fp8

1,386–35,385.02 ms

All latency percentiles 3 endpoints

Endpoint	P50	P75	P90	P99
DeepInfradeepinfra/fp8	1,358.5 ms	2,042 ms	2,904 ms	9,657.51 ms
GMICloudgmicloud/fp8	2,476.5 ms	2,819 ms	3,237.4 ms	6,127.76 ms
SiliconFlowsiliconflow/fp8	1,386 ms	1,840 ms	2,625.8 ms	35,385.02 ms

Output throughput · tokens per second

Throughput distributionRolling 30-minute snapshot · 3 endpoints

P50P75P90P99

DeepInfradeepinfra/fp8

86–205.34 t/s

GMICloudgmicloud/fp8

32–91.7 t/s

SiliconFlowsiliconflow/fp8

77–245.92 t/s

All throughput percentiles 3 endpoints

Endpoint	P50	P75	P90	P99
DeepInfradeepinfra/fp8	86 t/s	121 t/s	145 t/s	205.34 t/s
GMICloudgmicloud/fp8	32 t/s	37 t/s	41 t/s	91.7 t/s
SiliconFlowsiliconflow/fp8	77 t/s	176 t/s	209 t/s	245.92 t/s

Percentiles describe the observed request distribution. These are snapshots of a rolling window, not a historical time series.

Uptime

Compare successful request rates across the latest 5-minute, 30-minute, and 24-hour observations.

Successful requests by window

Availability by endpointCurrent overlapping windows · 3 endpoints

Endpoint5 minutes30 minutes24 hours

DeepInfradeepinfra/fp8Observed Oct 10, 2026, 11:31 AM UTC

100%

GMICloudgmicloud/fp8Observed Oct 10, 2026, 11:31 AM UTC

100%

99.92%

SiliconFlowsiliconflow/fp8Observed Oct 10, 2026, 11:31 AM UTC

100%

Reported by OpenRouter at the time of collection; rate-limited requests are excluded. A dash means no measurement was reported. This is not a live availability check.

Benchmarks

Review attributed evaluation results on each source's original scale.

External evaluations

deepseek/deepseek-v4-flash-vision-expObserved Oct 10, 2026, 11:31 AM UTC

Benchmark profileArtificial Analysis · shared source scale from 0 to 70

Coding

Agentic

47.5

Intelligence

34.8

Sources: Artificial Analysis and Design Arena, via OpenRouter. Scores retain their original scales and are not OpenAI Suite ratings.

Detailed evaluations

Evaluation scoresAccuracy or item F1 · higher is better

gpqa diamondDeepSeek: DeepSeek V4 Flash Vision Exp

88.019%

tau bench verified airlineDeepSeek: DeepSeek V4 Flash Vision Exp

74.8%

Evaluation	Configuration	Source	Results	Details
Composite indices	DeepSeek V4 Flash Vision (Max)deepseek/deepseek-v4-flash-vision-exp-20260821	Artificial Analysis	Intelligence 34.8 · Coding 65 · Agentic 47.5	Details
gpqa diamond	DeepSeek: DeepSeek V4 Flash Vision Expdeepseek/deepseek-v4-flash-vision-exp-20260821	OpenRouter	88.019%	Details
tau bench verified airline	DeepSeek: DeepSeek V4 Flash Vision Expdeepseek/deepseek-v4-flash-vision-exp-20260821	OpenRouter	74.8%	Details

Sources: Artificial Analysis, Design Arena and OpenRouter. Evaluation configurations and scales differ; results are not interchangeable. Updated Oct 24, 2026.

Apps & session costs

Compare observed median session costs for applications using this model, grouped by conversation length.

30-day sample · through Oct 18, 2026

Application	Model route	Turns	Median cost / session	Details
Hermes Agent	deepseek/deepseek-v4-flash-vision-exp-20260821	2–9 turns	$0.008629	Details
Hermes Agent	deepseek/deepseek-v4-flash-vision-exp-20260821	10–49 turns	$0.078946	Details

Source: OpenRouter, as of Oct 20, 2026. Published under CC BY 4.0. These are observed costs per session, not token prices or a forecast for your workload. Apps are compared separately.

Activity

Daily prompt and completion tokens reported for this model's routes in OpenRouter's top 50. Missing days have no published value.

Last 90 complete UTC days

Daily token usage1,093,742,983,026 reported tokens · 20 days with observations

88B0

2026-08-23
29,191,269,130 tokens

2026-08-25
32,630,656,289 tokens

2026-08-26
32,977,371,750 tokens

2026-08-27
31,602,924,117 tokens

2026-08-28
30,474,484,301 tokens

2026-08-29
25,280,994,735 tokens

2026-08-30
27,159,696,871 tokens

2026-08-31
33,714,811,394 tokens

2026-09-01
46,082,638,610 tokens

2026-09-02
79,145,189,963 tokens

2026-09-03
69,192,789,823 tokens

2026-09-04
74,504,701,007 tokens

2026-09-05
67,022,895,262 tokens

2026-09-06
85,022,814,346 tokens

2026-09-07
88,427,249,829 tokens

2026-09-08
81,173,771,131 tokens

2026-09-09
79,765,418,615 tokens

2026-09-10
76,250,648,361 tokens

2026-09-11
67,844,892,594 tokens

2026-09-12
36,277,764,898 tokens

Jul 26, 2026Oct 23, 2026

Daily observations · 20 days

Date (UTC)	Reported tokens	Reported routes
Sep 12, 2026	36,277,764,898	1
Sep 11, 2026	67,844,892,594	1
Sep 10, 2026	76,250,648,361	1
Sep 9, 2026	79,765,418,615	1
Sep 8, 2026	81,173,771,131	1
Sep 7, 2026	88,427,249,829	1
Sep 6, 2026	85,022,814,346	1
Sep 5, 2026	67,022,895,262	1
Sep 4, 2026	74,504,701,007	1
Sep 3, 2026	69,192,789,823	1
Sep 2, 2026	79,145,189,963	1
Sep 1, 2026	46,082,638,610	1
Aug 31, 2026	33,714,811,394	1
Aug 30, 2026	27,159,696,871	1
Aug 29, 2026	25,280,994,735	1
Aug 28, 2026	30,474,484,301	1
Aug 27, 2026	31,602,924,117	1
Aug 26, 2026	32,977,371,750	1
Aug 25, 2026	32,630,656,289	1
Aug 23, 2026	29,191,269,130	1

Source: OpenRouter (openrouter.ai/rankings), as of Oct 24, 2026. CC BY 4.0. Totals include only individually published routes. Tokenizers differ by provider; the aggregated “other” category is never assigned to a model.

Technical details & API

Explore specifications, supported parameters, reasoning controls, and API identifiers for each OpenRouter route.

Current route specifications

deepseek/deepseek-v4-flash-vision-expObserved Oct 10, 2026, 11:31 AM UTC

Context: 1,048,576
Maximum output: 262,144
Tokenizer: DeepSeek
Input: text, image
Output: text
Moderation: No
Added to OpenRouter: 2026-08-21

Base URLhttps://openrouter.ai/api/v1

Model IDdeepseek/deepseek-v4-flash-vision-exp

Version IDdeepseek/deepseek-v4-flash-vision-exp-20260821

Weights IDdeepseek-ai/DeepSeek-V4-Flash-Vision-Exp

Reasoning

Required: No
Enabled by default: Yes
Default effort: high
Supported efforts: max, high, low

Supported parameters

frequency_penalty
include_reasoning
logit_bias
logprobs
max_tokens
min_p
presence_penalty
reasoning
reasoning_effort
repetition_penalty
response_format
seed
stop
structured_outputs
temperature
tool_choice
tools
top_k
top_logprobs
top_p

Frequently asked questions

Answers based on the model metadata and serving offers currently in the catalog.

How much does DeepSeek V4 Flash Vision Exp cost?

Listed rates start at $0.2156 per million input tokens and $0.6468 per million output tokens. Prices vary by provider and configuration; see Pricing for details.

Which providers offer DeepSeek V4 Flash Vision Exp?

There are 18 cataloged offers across 15 providers. Compare model IDs, prices and limits in Providers.

What is the context window?

The cataloged model context is 1,000,000 tokens. Each serving endpoint may apply a different limit.

Does it support tools and structured output?

Tool calling: Not reported. Structured output: Not reported. Support can vary by endpoint.

Resources

MIT

Hugging Face ↗

Catalog history · 4 recorded revisions

Metadata observations since this model was first cataloged. These are not model release versions.

Oct 20, 2026 · 182512857de8
Sep 19, 2026 · 7e8b1dfbdf65
Sep 14, 2026 · 02f92315d598
Sep 12, 2026 · 87b04db1a84b