Back to AI Models

DeepSeek V4 Flash Vision (exp)

deepseek-v4-flash-vision-exp
Try with NaraRouter

An experimental vision-enhanced build of DeepSeek V4 Flash served via Fireworks, adding multimodal image understanding to the otherwise text-first efficient V4 Flash model.

TextVisionDeepseekDiscount 80%
Overview

Best for

MultimodalCodingGeneral assistance

Characteristics

FastStreamingMultimodal

Modalities

Text

Use cases

General AssistantCodingMultimodal Assistant
Endpoint Compatibility

Anthropic

Yes

Chat Completions

Recommended

Responses Api

Yes

Specification

Provider

DeepSeek

Context Window

1M

Available Access

Deepseek
Pricing

Prices shown per 1M tokens.

Input

Rp. 782

$0.04 USD

Output

Rp. 2.345

$0.13 USD

Cache

Rp. 65

$0 USD

All prices are estimates and subject to change.

Capabilities
ReasoningCodingTool CallingVisionStructured OutputStreaming

Similar Models

Related models selected by the catalog administrator.

DeepSeek v4 Flash
deepseek-v4-flash

DeepSeek's V4 Flash model, a fast, cost-efficient tier of the DeepSeek-V4 generation focused on general chat and coding with a 1M-token context; served as a text-only open-weights-style model.

Text
Claude Sonnet 5
claude-sonnet-5

Anthropic's Claude Sonnet 5, the balanced workhorse tier of the Claude family (between Opus and Haiku), tuned for strong agentic coding and everyday reasoning with 1M-token context and vision.

TextVision
GLM 5.3
glm-5.3

GLM-5.3 is Z.ai's flagship frontier model family, built with intensive reinforcement learning for major gains in coding, multi-step agents and emerging cybersecurity capability, and widely used for AI-assisted coding in harnesses like Codex and ZCode. It is the most advanced iteration of the GLM-5 line.

Text
DeepSeek V4 Flash Vision (exp) | NaraRouter