impulse

DeepSeek V4 Flash Vision Exp

Open weightsPreview

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,…

API id deepseek-v4-flash-vision-expReleased 21 Aug 2026 (Aug 2026)
Input / 1M
$0.30
at DeepSeek
Output / 1M
$1.20
at DeepSeek
Context window
1M
tokens
Max output
393K
tokens per reply
Hosts
5
7 listings
Released
Aug 2026
21 Aug 2026

Pricing

List prices at DeepSeek, in US dollars per million tokens.

TierInput / 1MOutput / 1M
Standard$0.30$1.20
Cached input
98% off input
$0.006
Cache write, 5 minutes$0.00

What it costs a month

At DeepSeek list prices, 30 days.

  • Chat assistant5,000 requests a day, 2K in / 0.5K out, cached$154
  • RAG over documents2,000 requests a day, 8K in / 0.6K out, cached$117
  • Coding agent1,000 requests a day, 40K in / 2K out, cached$185
  • Overnight batch20,000 requests a day, 4K in / 0.4K out$1,008
Use your own numbers

Where to run it

5 hosts serve DeepSeek V4 Flash Vision Exp. The comparison is on blended price, 3 parts input to 1 part output.

HostIn / out per 1M
DeepSeekLab API
deepseek-v4-flash-vision-exp
$0.30 / $1.20
Fireworks AIInference
deepseek-v4-flash-vision-exp
$0.22 / $0.66
−37% vs headline
deepseek-v4-flash-vision-exp
$0.389 / $1.17
+11% vs headline
GMI CloudInference
deepseek-v4-flash-vision-exp
$0.44 / $1.32
+26% vs headline
Novita AIInference
deepseek-v4-flash-vision-exp
$0.44 / $1.32
+26% vs headline
Show 2 more listings: regions, snapshots and variants
DeepSeek
Alias
deepseek-v4-flash-vision-exp
$0.30 / $1.20
Fireworks AI
Alias
deepseek-v4-flash-vision-exp
$0.22 / $0.66
−37% vs headline

Capabilities

Core

  • Streaming
  • Function calling
  • JSON mode
  • Structured outputs

Multimodal

  • Vision
  • Image input
  • Audio
  • Image generation
  • Video

Advanced

  • Reasoning
  • Tool calling
  • MCP
  • Prompt caching
  • Batch API
  • Embeddings
Takes in
TextImages
Produces
Text
Features
Assistant prefillParallel tool callsPrompt caching
Open weights
License: DeepSeek License (MIT code)
Runs in about 731.2 GB of GPU memory
Runtimes: SGLang, TGI, vLLM

Cheaper alternatives

Current models that cost less and keep most of what this one can do: tools, vision, reasoning, structured output, caching and enough context.

Same model, cheaper host

  • Fireworks AI
    deepseek-v4-flash-vision-exp
    $0.22 / $0.66
    37% less
  • Fireworks AI
    deepseek-v4-flash-vision-exp
    $0.22 / $0.66
    37% less

Smaller DeepSeek models

No cheaper DeepSeek model keeps its key capabilities.

Other frontier labs

  • GPT-5.4 Nano
    OpenAI · on OpenAI
    $0.20 / $1.25
    Has every key capability
    12% less
  • Claude Haiku 5.5
    Anthropic · on Anthropic
    $0.10 / $0.50
    Has every key capability
    62% less
  • GPT-6 Luna
    OpenAI · on OpenAI
    $0.10 / $0.50
    Has every key capability
    62% less
  • GPT-4o Mini
    OpenAI · on OpenAI
    $0.15 / $0.60
    Lacks reasoning
    50% less

Open-weight picks

  • Qwen3.7 Plus
    Alibaba Qwen · on Replicate
    $0.276 / $1.10
    Has every key capability
    8% less
  • MiMo V2.6 Pro
    Xiaomi · on GMI Cloud
    $0.348 / $0.696
    Has every key capability
    17% less
  • Muse Glimmer 30B
    Meta · on EmpirioLabs AI
    $0.20 / $0.80
    Has every key capability
    33% less
  • Nemotron 3.5 Super VL
    NVIDIA · on Kyma API
    $0.081 / $0.472
    Has every key capability
    66% less

Call it

Using DeepSeek's API.

Python
from litellm import completion

response = completion(
    model="deepseek/deepseek-v4-flash-vision-exp",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Quick answers

How much does DeepSeek V4 Flash Vision Exp cost?

On DeepSeek, DeepSeek V4 Flash Vision Exp costs $0.30 per million input tokens and $1.20 per million output tokens. Cached input costs $0.006 per million tokens. The cheapest host is Fireworks AI, at $0.22 per million input tokens and $0.66 per million output tokens.

Which APIs offer DeepSeek V4 Flash Vision Exp?

DeepSeek V4 Flash Vision Exp runs on 5 APIs: DeepSeek, Fireworks AI, Kyma API, GMI Cloud and Novita AI.

What is DeepSeek V4 Flash Vision Exp's context window?

1M tokens, with up to 393K tokens in one reply.

Is DeepSeek V4 Flash Vision Exp being retired?

No retirement date is scheduled for it on any API we track.

What are cheaper alternatives to DeepSeek V4 Flash Vision Exp?

Models that cost less and keep most of what it can do: GPT-5.4 Nano on OpenAI (12% less), Claude Haiku 5.5 on Anthropic (62% less) and Qwen3.7 Plus on Replicate (8% less). DeepSeek V4 Flash Vision Exp itself costs 37% less on Fireworks AI. Savings are on a mix of 3 input tokens to 1 output token.

When was DeepSeek V4 Flash Vision Exp released?

DeepSeek released it on 21 Aug 2026.

What is DeepSeek V4 Flash Vision Exp's API model id?

Use deepseek-v4-flash-vision-exp with DeepSeek.

Sources

Raw catalog entry for deepseek-v4-flash-vision-exp on DeepSeek
{
  "cache_creation_input_token_cost": 0,
  "cache_read_input_token_cost": 6e-9,
  "input_cost_per_token": 3e-7,
  "input_cost_per_token_cache_hit": 6e-9,
  "litellm_provider": "deepseek",
  "max_input_tokens": 1000000,
  "max_output_tokens": 393216,
  "mode": "chat",
  "off_peak_pricing": {
    "cache_read_input_token_cost": 3e-9,
    "input_cost_per_token": 1.5e-7,
    "output_cost_per_token": 6e-7,
    "windows": [
      {
        "hours_utc": [
          "00:00-01:00",
          "04:00-06:00",
          "10:00-00:00"
        ],
        "weekdays": [
          1,
          2,
          3,
          4,
          5
        ]
      },
      {
        "hours_utc": "00:00-00:00",
        "weekdays": [
          6,
          7
        ]
      }
    ]
  },
  "output_cost_per_token": 0.0000012,
  "source": "https://api-docs.deepseek.com/quick_start/pricing",
  "supported_endpoints": [
    "/v1/chat/completions"
  ],
  "supports_assistant_prefill": true,
  "supports_function_calling": true,
  "supports_native_streaming": true,
  "supports_parallel_function_calling": true,
  "supports_prompt_caching": true,
  "supports_reasoning": true,
  "supports_response_schema": true,
  "supports_system_messages": true,
  "supports_tool_choice": true,
  "supports_vision": true
}