Tasks and inputs
DeepSeek V4.1 Flash is listed for Image-text-to-text, Text generation. It accepts text, image and produces text. Use the publisher’s input format and preprocessing for your chosen deployment.
SyncDev model directory · Profile updated Oct 13, 2026
DeepSeek V4.1 Flash is a multimodal mixture-of-experts model released, with native image understanding.
The specification score summarizes recorded features and limits. It is not a benchmark result.
Pricing conditions: Off-peak USD rates per million tokens are $0.15 uncached input, $0.003 cached input and $0.60 output. Peak rates are twice those amounts, during 01:00–04:00 and 06:00–10:00 UTC on weekdays excluding Chinese public holidays. The current model ID is deepseek-flash.
Recorded modality support shows what this model can accept and produce. An undocumented modality is shown as unknown rather than assumed to be unsupported.
Accepts
Produces
DeepSeek V4.1 Flash is a multimodal mixture-of-experts model released, with native image understanding.
The current Flash API name is deepseek-flash. The older deepseek-v4-flash and deepseek-v4-flash-vision-exp names are compatibility routes to V4.1 Flash, so they should not be used as proof that the retired models still serve requests. An application migrating from V4 should record both the requested identifier and the model now behind it.
DeepSeek describes 552B total parameters, with 8B active for input encoding and 16B active for output decoding. A single active-parameter figure would conceal that asymmetry. This profile leaves that combined field unset. The published weights use MIT licensing; serving a checkpoint still requires compatible inference support and an appropriate deployment.
Use screenshots, charts and documents from the finished application. Include tiny labels, incomplete screenshots and images that contradict the written prompt. Check whether the response separates visible evidence from inference. For an agent, test tool arguments, recovery after an error and whether it respects the supplied task boundaries.
DeepSeek applies peak and off-peak rates. A stored single token price cannot describe both schedules. Confirm the current pricing page and the actual host before budgeting. Earlier launch announcements and later service notices can disagree about an older endpoint; use the current API reference for the route you are calling.
Publisher source 1 · Publisher source 2 · Publisher source 3 Deployment limits and commercial terms belong to the selected checkpoint or API route.
Compare the model on the work you need to complete.
DeepSeek V4.1 Flash is listed for Image-text-to-text, Text generation. It accepts text, image and produces text. Use the publisher’s input format and preprocessing for your chosen deployment.
Documented limits: 1.05M context and 393.22K maximum output. The listed route has starting rates of $0.15 input and $0.60 output per million tokens. Measure a completed task using the prompts, documents and tools your application will use. Include reasoning and retries in the result.
The weights are downloadable under MIT. Check runtime support and size the complete model and working context for your expected concurrency.
These records share a directory category. Check the source card, output type and deployment requirements before treating a model as a replacement.
Anthropic
Claude’s small model for high-volume extraction, classification and routing, with adaptive thinking and a 1M-token context window.
Mistral AI
Mistral Large 4 is an open-weight multimodal model released in public preview on the publisher’s announced release, for general reasoning, coding and agent applications.
OpenAI
GPT-6.1 Sol is OpenAI's updated Sol model for coding and professional tasks. It accepts text and images and returns text.
inclusionAI
Ling 3.1 Flash is inclusionAI’s hybrid reasoning model for coding and tool agents, available through a hosted AI Gateway route.
Anthropic
Claude Sonnet 5.5 is an Anthropic language model that accepts text and images and returns text.
MiniMax
MiniMax M3.1 Flash Preview is a multimodal reasoning model for code and agent workflows, accessed through eligible MiniMax plans and tools.