DeepSeek-V4-Flash-Vision-Exp

Specs

Input price

$0.14/M

Output price

$0.28/M

Context window

1M tokens

Capabilities

• Text-plus-image input at Flash-class pricing, MIT-licensed weights
• Vision-aware agents: Chartography 64.3, Agents' Last Exam 27.3
• Text agentic performance held up vs Flash-0731 (Terminal-Bench 83.9)
• DSpark speculative decoding and a 1M-token context window

Best for

• Document-heavy pipelines: scans, screenshots, charts, and photos
• Statistical chart reading for lab trends and utilization dashboards
• High-volume multimodal extraction where closed vision APIs are overkill

Limitations to keep in mind

• Experimental checkpoint - can be superseded quickly on a fast-moving family
• Not a radiology or pathology tool; reads documents, not diagnostics
• Latency under heavy image load not independently benchmarked

HIPAA-compliant hosting

DeepSeek-V4-Flash-Vision-Exp is available under our HIPAA-compliant design-partner program, with a signed Business Associate Agreement (BAA), encryption in transit and at rest, and access controls. We are currently onboarding design partners, with general availability coming soon. Run the model behind a unified OpenRouter-style API and swap to another model with a single parameter.

Pricing in context

Open source models like DeepSeek-V4-Flash-Vision-Exp can be served efficiently on optimized inference infrastructure, with savings passed through to you. Exact savings depend on the model and your volume, but open source inference is typically a fraction of the per-token cost of closed-source frontier models on Azure OpenAI or AWS Bedrock - without cloud egress lock-in or minimum commitments.