Skip to content
Alpha
  • Models
  • Rankings
  • Ori

Footer

OpenRouterOpenRouter
© 2026 OpenRouter, Inc

Product

  • Chat
  • Rankings
  • Benchmarks
  • Apps
  • Discover
  • Models
  • Ori
  • Collections
  • Providers
  • Tools
  • Pricing
  • Business
  • EU routing
  • Enterprise
  • Labs

Company

  • About
  • Blog
  • Careers
    Hiring
  • Privacy
  • Terms of Service
  • Trust Center
  • Support
  • Works With OR
  • Data
  • Brand

Developer

  • Documentation
  • API Reference
  • Developer Platform
  • Status
  • AI Site Map

Connect

  • Discord
  • GitHub
  • LinkedIn
  • X
  • YouTube
Favicon for deepseek

DeepSeek: DeepSeek V4.1 Flash

deepseek/deepseek-v4.1-flash

Model weights
Compare

DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the first built on the company's Causal Encoder-Decoder (CED) architecture. It activates 8B parameters on input and 16B on output from a 552B-parameter backbone, an asymmetric split that keeps per-token compute low relative to the model's total size. Image understanding is native to the architecture, with visual and text embeddings trained jointly from the start of pre-training rather than added afterward as in the earlier experimental V4 Flash Vision ExpOpens in new tab.

It is suited for coding, terminal, and computer-use agents, along with long-horizon tasks that must run to completion across many steps and long-context analysis. Compressed KV caching cuts cache memory to roughly a quarter of the previous Flash generation, significantly reducing costs on agentic workloads. DeepSeek positions it as the cost-efficient tier of the V4.1 family and reports that it exceeds V4 ProOpens in new tab on performance, speed, and task completion time.

Modalities
In / Out Price
$0.02 / $1per 1M
Context
1.0M
Released
Sep 10, 2026
Compare
ProvidersPricingPerformanceUptimeAppsActivityDocumentationFAQExplore

Providers

Different companies host the same model. OpenRouter routes your request to one of them based on the routing mode you pick — Balanced (price + speed), Nitro (fastest), or Floor (cheapest).

Pricing

The average price customers actually pay for this model, next to the prices providers post. Caching and discounts mean the price actually paid is often well below the listed one.

Performance

Throughput is how fast the model writes (tokens per second — higher is better). Latency is total round-trip time (lower is better). TTFT is time-to-first-token — how long before you see anything appear (lower is better).

Uptime

Uptime is the percentage of the past 3 days that at least one provider was responding to requests. Availability is the percentage of time that inference was successfully served. OpenRouter continuously monitors and uses the next-best provider when one returns an error.

Apps

Public apps that send the most traffic to this model. Good signal for what real production workloads look like — and a hint at which use cases this model is best suited for.

Activity

Token volume and request traffic to this model over time.

API

Drop-in code to call this model. OpenRouter's API is OpenAI-compatible — most SDKs work by just swapping the base URL. The only thing that changes between models is the model slug below.

DeepSeek: DeepSeek Flash versions

Not currently available through EU in-region routing. In-region routing docs

Documentation

Guides, cookbooks, blog posts, and tools for building with DeepSeek: DeepSeek V4.1 Flash on OpenRouter.

  • Does DeepSeek V4 Have Vision?Blog post

Explore more models

AI Models with Vision: Multimodal LLMs for Image UnderstandingCollectionAI Model RankingsRanking
Latency / throughput
$0.070$0.500$0.008000.57s131 tps
100.00%
$0.190$0.650$0.012000.35s258 tps
100.00%
$0.165$0.660$0.006000.80s183 tps
100.00%
$0.210$0.840$0.004200.50s228 tps
100.00%
$0.150$0.650$0.002000.90s161 tps
99.85%
$0.100$0.600$0.010000.52s226 tps
100.00%
$0.300$1.200$0.007000.42s282 tps
100.00%
$0.170$0.670$0.006000.41s141 tps
99.59%
$0.103$0.645$0.008000.31s121 tps
100.00%
$0.103$0.645$0.008000.32s114 tps
100.00%
$0.300$1.200$0.003000.49s234 tps
99.88%
30% off
$0.20$0.140$0.60$0.420$0.006$0.004200.65s71 tps
100.00%
$0.300$1.200$0.007500.37s255 tps
100.00%
$0.300$1.200$0.006000.81s158 tps
100.00%
$0.290$1.200$0.007000.49s121 tps
98.10%
$0.135$0.540$0.013500.87s71 tps
100.00%
$0.300$1.200$0.006000.27s279 tps
100.00%
51% off
$0.30$0.147$1.20$0.588$0.006$0.002941.52s78 tps
100.00%
30% off
$0.15$0.105$1.20$0.840$0.01$0.007000.31s114 tps
97.45%
$0.260$1.170$0.005800.69s79 tps
99.94%
$0.300$1.200$0.006000.41s110 tps
100.00%
30% off
$0.30$0.210$1.20$0.840$0.006$0.004202.56s76 tps
99.61%
35% off
$0.30$0.195$1.20$0.780$0.006$0.003901.40s82 tps
93.21%
$0.020$1.000$0.015001.05s84 tps
99.71%
$0.090$0.250$0.006000.66s58 tps
100.00%
$0.300$1.200$0.010000.66s154 tps
98.44%
$0.300$1.200$0.006001.16s64 tps
100.00%
$0.300$1.200$0.006002.08s85 tps
100.00%
$0.330$1.320$0.033001.34s79 tps
99.94%
$0.200$1.000--1.16s38 tps
89.43%
Priority30% off
$0.30$0.210$0.90$0.630$0.009$0.006300.71s85 tps
48.63%
40% off
$0.30$0.180$1.20$0.720$0.006$0.003601.97s79 tps
99.97%
$0.300$1.200$0.030000.41s91 tps
100.00%
$0.300$1.200$0.030000.88s68 tps
98.48%
$0.090$0.250$0.006000.45s49 tps
100.00%
$0.300$1.200$0.006000.94s79 tps
99.97%
$0.300$1.200$0.007000.29s255 tps
100.00%
Fast
$0.600$2.400$0.014000.44s161 tps
99.71%
$0.090$0.180$0.018003.77s83 tps
93.56%
Not used in Standard routing:Why these endpoints are not used
Flex30% off
$0.16$0.112$0.48$0.336$0.0048$0.003363.53s27 tps
97.71%
$0.450$1.800$0.009000.24s173 tps
100.00%
Throughput
282tok/s

P50, best across providers

Latency
0.24s

P50, best provider

AutoExacto benchmarks
GPQA DiamondTAU-BenchGoogle Vertex90.1%--NextBit91.0%76.2%DigitalOcean89.8%77.4%StreamLake90.5%76.0%NovitaAI90.0%76.4%
+35 more providers
Uptime (3d)The model was reachable. Request routed to a provider.
100.00%
Availability (3d)The model returned inference from any provider. Errors and empty responses count against it.
99.90%

Availability over the last 3 days

Last 72 hours
Availability 99.90%
3 Days Ago2 Days AgoYesterdayNow

Availability over the last 24 hours

OpenRouter Availability
99.91%
Without Routing
87.84%

When an error occurs in an upstream provider, we can recover by routing to another healthy provider, if your request filters allow it. You can access per-provider uptime data programmatically through the Endpoints API. Learn more about our load balancing and provider routing options.

1.
Favicon for https://nousresearch.com
Hermes Agent
Hermes Agent is an open-source, self-improving AI agent by Nous Research that runs persistently with memory across sessions, and builds reusable skills from experience. It comes with 40+ built-in tools, including web search, browser automation, and vision, plus scheduled automations and subagents.
12Ttokens
2.
Favicon for https://claude.ai/apple-touch-icon.png
Claude Code
Claude Code is Anthropic's agentic coding tool that reads your entire codebase, plans and executes changes across files, runs tests, and iterates on failures, all from natural language prompts.
5.37Ttokens
3.
Favicon for https://pi.dev/
pi
There are many coding agents, but this one is yours.
4.46Ttokens
4.
Favicon for https://github.com/deepseek-ai/deepseek-harness
DeepSeek Harness
new
2.87Ttokens
5.
Favicon for https://docs.all-hands.dev/
OpenHands
AI coding agent that can run commands, browse the web, call APIs
2.08Ttokens

Frequently asked questions

DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the first built on the company's Causal Encoder-Decoder (CED) architecture. It activates 8B parameters on input and 16B on output from a 552B-parameter backbone, an asymmetric split that keeps per-token compute low relative to the model's total size.

The DeepSeek V4.1 Flash documentation on OpenRouter includes the Does DeepSeek V4 Have Vision? blog post. The quickstart covers API keys and your first request, and the API reference lists every parameter.

DeepSeek V4.1 Flash costs $0.02/M input tokens and $1.00/M output tokens, with separate rates for Cache Read at $0.015/M tokens.

DeepSeek V4.1 Flash has a 1,048,576 token context window. It supports up to 1,048,576 completion tokens.

Yes. DeepSeek V4.1 Flash accepts tools and tool_choice for function calling. It also supports structured outputs via a JSON schema in response_format.

DeepSeek V4.1 Flash accepts text and images as input and returns text.

DeepSeek V4.1 Flash is served by 33 providers on OpenRouter: Open Inference, Relace, Decart, Sail Research, Morph, inference.net, Reka AI, DeepInfra and 25 more. Requests are routed to the best available provider, with automatic failover to the others, and you can pin or exclude providers with provider routing.

DeepSeek V4.1 Flash was released on September 10, 2026.