Bertrande

Catalogue / OpenRouter / DeepSeek: DeepSeek V4.1 Flash (sold by OpenRouter)

DeepSeek: DeepSeek V4.1 Flash (sold by OpenRouter)

Prices and limits, each value carrying the date it was observed and the source it came from.

As of 28 October 2026, DeepSeek V4.1 Flash costs $0.045 per million input tokens and $0.500 per million output tokens on OpenRouter. Observed on OpenRouter's own published price list on 28 October 2026. This is OpenRouter's resale price, not the model vendor's list price.

Input $0.045 per 1M
Output $0.500 per 1M
Context 1 048 576 tokens

What does DeepSeek V4.1 Flash cost?

ParameterValueUnit ObservedSourceProof
Input price $0.045 USD per 1M tokens 28 October 2026
checked 3 November 2026
official API L1 648885296329
Output price $0.500 USD per 1M tokens 28 October 2026
checked 3 November 2026
official API L1 648885296329
Context window 1 048 576 tokens 30 October 2026
checked 3 November 2026
official API L1 7b501cd7e861
Maximum output 943 718 tokens 28 October 2026
checked 3 November 2026
official API L1 648885296329
Intelligence index (Artificial Analysis) 24.7 index 0-100 24 October 2026 recheck due official API L4 0ea5cd4fefac
Input modalities
  • image
  • text
list 30 October 2026
checked 3 November 2026
official API L1 7b501cd7e861
Output modalities
  • text
list 30 October 2026
checked 3 November 2026
official API L1 7b501cd7e861
Cache read price $0.006 USD per 1M tokens 28 October 2026
checked 3 November 2026
official API L1 648885296329
Listed since 2026-09-10 date 30 October 2026
checked 3 November 2026
official API L1 7b501cd7e861
Supported parameters
  • frequency_penalty
  • include_reasoning
  • logit_bias
  • logprobs
  • max_tokens
  • min_p
  • parallel_tool_calls
  • presence_penalty
  • reasoning
  • reasoning_effort
  • repetition_penalty
  • response_format
  • seed
  • stop
  • structured_outputs
  • temperature
  • tool_choice
  • tools
  • top_a
  • top_k
  • top_logprobs
  • top_p
list 31 October 2026
checked 3 November 2026
official API L1 aa4b1be8e631

Does the price change with the hour or the size of the request?

Yes. This provider publishes conditional rates, so the figure above is not the only price you can pay. Each rate below applies under a stated condition, taken verbatim from the provider's own catalogue.

ParameterRate AppliesObserved
Input price $0.150 weekdays 00:00 to 01:00 UTC; weekdays 04:00 to 06:00 UTC; weekdays 10:00 to 00:00 UTC; weekends (all day, UTC) 10 September 2026
Input price $0.300 weekdays 01:00 to 04:00 UTC; weekdays 06:00 to 10:00 UTC 12 October 2026
Output price $0.600 weekdays 00:00 to 01:00 UTC; weekdays 04:00 to 06:00 UTC; weekdays 10:00 to 00:00 UTC; weekends (all day, UTC) 10 September 2026
Output price $1.20 weekdays 01:00 to 04:00 UTC; weekdays 06:00 to 10:00 UTC 12 October 2026
Cache read price $0.003 weekdays 00:00 to 01:00 UTC; weekdays 04:00 to 06:00 UTC; weekdays 10:00 to 00:00 UTC; weekends (all day, UTC) 10 September 2026
Cache read price $0.006 weekdays 01:00 to 04:00 UTC; weekdays 06:00 to 10:00 UTC 12 October 2026

We record each conditional rate as its own dated observation, so a discount window that disappears shows up as a change like any other.

How reliable is each host behind DeepSeek V4.1 Flash?

The cheapest offer is not a bargain if it is down. OpenRouter routes traffic to several upstream hosts for this model and publishes the availability its own router recorded over the last day for each of them. Least reliable first.

HostAvailability, last day Observed
baseten/fast 25.75 % 31 October 2026
reka/fp4 88.91 % 17 September 2026
fireworks 96.75 % 3 November 2026
google-vertex/global 97.24 % 3 November 2026
reka 98.39 % 3 November 2026
wafer 98.75 % 15 October 2026
dekallm 98.80 % 12 October 2026
io-net/fp8 98.80 % 2 October 2026
gmicloud/fp8 98.90 % 3 November 2026
makora 98.93 % 23 September 2026
makora/fp8 99.15 % 3 November 2026
krea/fp8 99.19 % 30 September 2026
decart/fp4 99.29 % 3 November 2026
inference-net 99.50 % 6 October 2026
fireworks/spot 99.51 % 3 November 2026
phala 99.65 % 3 November 2026
morph 99.76 % 26 September 2026
baseten/us 99.80 % 3 November 2026
mancer/fp8 99.85 % 3 November 2026
cloudflare 99.86 % 17 October 2026
morph/fp8 99.86 % 3 November 2026
open-inference/fp4 99.86 % 3 November 2026
relace/fp4 99.86 % 3 November 2026
baseten/fp32 99.86 % 16 October 2026
venice/fp8 99.93 % 3 November 2026
nextbit/fp8 99.94 % 6 October 2026
alibaba 100.00 % 2 November 2026
atlas-cloud/fp8 100.00 % 30 October 2026
baidu/fp8 100.00 % 2 November 2026
baseten 100.00 % 3 November 2026
baseten/fp8 100.00 % 15 October 2026
cohere 100.00 % 31 October 2026
coreweave/fp8 100.00 % 3 November 2026
crusoe/fp8 100.00 % 31 October 2026
deepinfra/fp8 100.00 % 31 October 2026
deepinfra/fp8/flex 100.00 % 2 November 2026
deepinfra/fp8/priority 100.00 % 1 November 2026
deepseek 100.00 % 17 October 2026
digitalocean 100.00 % 29 October 2026
fireworks/us 100.00 % 2 November 2026
inference-net/fp8 100.00 % 29 October 2026
inference-net/us 100.00 % 29 October 2026
ionstream 100.00 % 22 October 2026
modal 100.00 % 3 November 2026
modelrun/fp4 100.00 % 26 October 2026
nebius/fp8 100.00 % 3 November 2026
novita/fp8 100.00 % 28 October 2026
parasail/fp8 100.00 % 29 October 2026
primeintellect 100.00 % 1 November 2026
relace 100.00 % 30 October 2026
relace/fast 100.00 % 3 November 2026
sail-research/fp4 100.00 % 3 November 2026
sail-research/fp4/us 100.00 % 27 October 2026
siliconflow/fp8 100.00 % 3 November 2026
streamlake/fp8 100.00 % 30 October 2026
together 100.00 % 30 October 2026
wafer/mxfp4 100.00 % 29 October 2026
wafer/us 100.00 % 29 October 2026

This is availability as OpenRouter's router observed it, not as you would measure it calling the host directly: the network path is not the same. We read it once a day and keep every reading, so a host that was unreliable last August can still be shown to have been unreliable last August.

We publish no time to first token and no throughput. OpenRouter's API carries fields for both and leaves them empty on every endpoint we have read, and we will not fill a number the source does not give.

How has DeepSeek V4.1 Flash moved?

Output price, 61 observations between 10 September 2026 and 28 October 2026. Range $0.0112 to $2.40. Drawn as a step: a value holds until the next observation, we do not interpolate between them.

How capable is DeepSeek V4.1 Flash?

IndexScoreObservedSource
Intelligence index (Artificial Analysis)24.7 24 October 2026 Artificial Analysis L4

These indices are produced by Artificial Analysis and relayed through the catalogue API. They are not our own measurement. See cost per point of capability to compare them against price.

Has DeepSeek V4.1 Flash changed price?

DetectedParameter FromToChange
24 October 2026 Supported parameters n/a n/a changed
22 October 2026 Supported parameters n/a n/a changed
17 October 2026 Intelligence index (Artificial Analysis) 39.5 24.7 -37.5%
10 September 2026 Supported parameters n/a n/a changed

What does DeepSeek V4.1 Flash cost elsewhere?

We track this model at 6 sellers. The same model rarely costs the same everywhere, and the difference is not visible from any single seller's catalogue.

Compare DeepSeek V4.1 Flash across 6 sellers, side by side, or see every model tracked at more than one seller.

Other services from OpenRouter

All OpenRouter services

Where these figures come from

Source: OpenRouter. Every figure on this page was read from OpenRouter's own published price list, and every row above links straight back to it. We publish factual values with attribution, we do not reproduce OpenRouter's content, and we send readers to them. If you publish this source and want us to stop tracking it, write to contact@bertrande.com and we will act on it without argument.

How is this page verified?

Every figure above comes from OpenRouter's own published price list, which we rank as a level 1 source. We re-check it every 6 hours. Each observation stores the exact source URL and a SHA-256 fingerprint of the source content at that moment, which is what lets us prove what the source said on a given day.

The Proof column shows the first twelve characters of the SHA-256 fingerprint of the source content at the moment we read it. Hover it for the full value. It is what lets us demonstrate later what the source said on a given day, after that page has changed or vanished. The full fingerprints are in data.json.

No value on this page was produced, estimated or completed by a language model. Where we do not have a verified figure, the row is absent rather than filled in. Read the full methodology, or report an error.

Take the data, or watch this page

This page is available as data.json (including the full observation history of every parameter) and data.csv, under CC-BY-4.0, free, with attribution. See all downloads.

Want to know when this changes? Point a feed reader or a script at the Atom feed, which carries every confirmed change and is rebuilt hourly, or poll this page's data.json, where every figure carries its observation date and the fingerprint of its source. Both work today and need no address.

Email alerts are open, and nothing is sent until you confirm from a link we email you. Subscribe, or see what an alert looks like.