Kyma APIKyma API
DeepSeek

DeepSeek

DeepSeek V4.1 Flash

#4 on Kyma

DeepSeek V4.1 Flash is the lab's current general model: a 1M-token context window, native image input and a thinking mode, at the Flash price band. DeepSeek positions it ahead of V4 Pro and routes V4 Pro requests onto it from 2026-09-14, so it is the DeepSeek id to pick for new work.

Modalities

Text+Image → Text

Below release

fp4

Weights

Published

deepseek-ai/DeepSeek-V4.1-Flash

Where it sits in the catalogue Kyma measures

Every number against every text model Kyma prices per token — a stated rule, not a chosen line-up.

Price · input#6/71

$0.1048

output $0.3141

median $1.013 · best $0.0494

Price · output#5/70

$0.3141

per 1M tokens generated

median $3.531 · best $0.218

Availability · 30d

Throughput

52 tok/s

from traffic, 7d — real prompts, not a probe

Response time

2.85 s

from traffic, 7d — real prompts, not a probe

Context

1.05M

max output 131,072

published by DeepSeek, not measured here

Tick above each rail is this model, below it the other 70. Dashed rule is the field median, solid is its best. Better is left; the axis stops at the 90th percentile, so a few models sit past its right edge.

Usage

How much this model is actually called here.

Rank

#4

of 101 active models

Tokens served

319.8M

all-time

Platform share

7.3%

of all tokens

Tokens · last 14 daysSep 26 → Oct 9

How it behaves under real clients

The same model answering different prompt shapes — measured, not benchmarked.

ClientRequestsTokensThroughputTo first tokenp95 totalCompleted
OpenClaw2,04329.6M53 tok/s0.75 s5.13 s100%
OpenCode1,723238.2M52 tok/s2.42 s21 s100%
Node.js App81170.2M85 tok/s1.02 s22 s100%
Java/Kotlin App40942.5M48 tok/s2.45 s21 s100%
Python App3682.5M36 tok/s3.89 s10 s100%
Cline35817.6M32 tok/s3.39 s21 s100%
OpenAI Python18812.5M58 tok/s1.46 s19 s100%

Grouped by the client that sent the request. Not a benchmark: the same model on different prompt shapes. Throughput and time-to-first-token are medians; p95 is the slowest response in twenty. 1 client under 100 requests not shown.

Pricing

Pay per token. Cached input bills at this model’s own cached rate, listed below.

$0.1048 /1M input$0.3141 /1M output
Hobby10 req/day · 2K in / 500 out
~$0.11/mo
Production1,000 req/day · 2K in / 500 out
~$11.00/mo
Scale20,000 req/day · 2K in / 500 out
~$220/mo
+ Estimate your workload
1,000
2,000
500

Estimated monthly cost

$11.00

$0.3667 / day on DeepSeek V4.1 Flash

Same workload on:

DeepSeek V4 Flash Vision$26.40+140%
DeepSeek V4 Flash$11.37+3%

Estimates use list pricing. Actual bills depend on real token counts, and every response includes its exact cost.

When to use DeepSeek V4.1 Flash

Updated 2026-09-12

Where this model earns its cost — and where it doesn't.

DeepSeek released V4.1 Flash on 2026-09-10 as the successor to both V4 Flash and V4 Pro. Its card lists a 1M context window, thinking and non-thinking modes, JSON output, tool calls and image input. The lab's own note says its results land ahead of V4 Pro and that its KV cache needs a quarter of the memory of the previous generation.

On Kyma it is served through the OpenAI-compatible endpoint with automatic failover across three routes and exact cost reporting in the usage.cost field. Prompt caching is not exposed on the primary route, so every input token bills at the input rate.

The weights are published under the MIT license. The output cap on Kyma is 131,072 tokens per request.

Coding

Tool calling and structured outputs with a 1M-token window put whole repositories in one request.

Vision at value pricing

Image input on a Flash-priced model — screenshots, diagrams and documents without moving to a flagship.

Long-context analysis

The 1M-token window takes document collections or long agent transcripts in a single call.

High-volume pipelines

Flash-band pricing suits background jobs and batch processing that run thousands of times a day.

Not ideal for: Prompt-cache-heavy workloads — the primary route publishes no cached-input rate, so repeated prefixes bill in full.

How it compares

Against the peers people actually weigh it against.

SpecDeepSeek V4.1 FlashDeepSeek V4 Flash VisionDeepSeek V4 Flash
Input /1M$0.1048$0.2514$0.1264
Output /1M$0.3141$0.7542$0.2526
Context1M1M1M
ToolsYesYesYes
ReasoningYesYesYes
Throughput52 tok/s14 tok/s32 tok/s

Quick start

Up and running in under two minutes.

  1. 1

    Create an API key

    Sign up and grab a key from the dashboard — $0.50 free credit on the free tier, which covers DeepSeek V4.1 Flash. No card required.

    Get API key →
  2. 2

    Make your first request

    Drop in your key and send a chat completion — fully OpenAI-compatible.

    curl https://kymaapi.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v4.1-flash",
    "messages": [
      {"role": "user", "content": "Explain prompt caching in one paragraph."}
    ]
  }'
  3. 3

    Stream responses

    Add "stream": true to receive tokens as they arrive.

    curl https://kymaapi.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v4.1-flash",
    "stream": true,
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

FAQ

Common questions about this model.

What is the context window of DeepSeek V4.1 Flash?

DeepSeek V4.1 Flash has a 1M-token context window — roughly 1542 pages of text in a single request.

How much does the DeepSeek V4.1 Flash API cost?

$0.1048 per 1M input tokens and $0.3141 per 1M output tokens. No subscription; you pay only for what you use.

Does DeepSeek V4.1 Flash support function calling?

Yes — DeepSeek V4.1 Flash supports tool/function calling and structured outputs (JSON mode), so it works with agent frameworks out of the box.

Are the weights for DeepSeek V4.1 Flash publicly available?

Yes. DeepSeek publishes DeepSeek V4.1 Flash's weights as deepseek-ai/DeepSeek-V4.1-Flash, so you can download and run the model yourself (https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash, read 2026-09-12). Kyma serves it because it is convenient and has failover behind it, not because it is the only way to reach it.

Is DeepSeek V4.1 Flash ever served below the precision its creator released it at?

Sometimes. DeepSeek V4.1 Flash was released by DeepSeek at fp8 (https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash, read 2026-09-12), and at least one route serving it here reports fp4 — narrower than that. Weights compressed below the release usually answer close to it, but it is not the identical artefact, and nothing in the response tells you which one answered. Kyma checked the routes on 2026-09-21, so it is your call rather than a silent one.

How do I use DeepSeek V4.1 Flash?

Kyma is OpenAI-compatible: point your SDK's base URL at https://kymaapi.com/v1, use your Kyma API key, and set the model to deepseek-v4.1-flash. Signing up is free and includes $0.50 of credit on the free tier, which covers this model — no card required.

How is DeepSeek V4.1 Flash different from V4 Flash?

It is a newer checkpoint (2026-09-10) with native image input and a thinking mode, and DeepSeek says its results land ahead of V4 Pro. V4 Flash stays on Kyma as the cheaper previous checkpoint.

Does it support images?

Yes — image input is native. Send images the same way as any other vision model on the OpenAI-compatible endpoint.

Why use DeepSeek V4.1 Flash through Kyma?

One API key and one OpenAI-compatible endpoint cover this and every other model on the platform, with automatic failover across its three routes and the exact cost of every request in usage.cost.

Start with $0.50 free credit on the free tier — no card required.Create account →

More models by DeepSeek

ModelContextInputOutput
DeepSeekDeepSeek V4 Flash Vision1M$0.2514$0.7542
DeepSeekDeepSeek V4 Flash1M$0.1264$0.2526
DeepSeekDeepSeek V3160K$0.518$0.7571
DeepSeekDeepSeek R164K$0.7425$2.957