DeepSeek R1 0528

May 28th update to the original DeepSeek R1 Performance on par with OpenAI o1, but open-sourced and with fully open reasoning tokens. It's 671B parameters in size, with 37B active in an inference pass.

Fully open-source model.

deepseek/deepseek-r1-0528

Organization: DeepSeek
Family: deepseek-thinking
Providers: 16
Context: 163,840
Output limit: 32,768
Knowledge: —
Release: 2025-05-28
Updated: —
Weights: Open
Input: text
Output: text
Capabilities: —

Providers

Compare every cataloged offer for this model, including its provider, route, token limits, price, and capabilities.

21 offers · USD per million tokens

Provider	Model ID	Context	Output	Input / output · 1M	Reasoning	Tools	Structured	Details
Deep Infra	deepseek-ai/DeepSeek-R1-0528	163,840	64,000	— / —	—	—	—	View
Eden AI	tensorx/deepseek/deepseek-r1-0528	164,000	32,768	— / —	—	—	—	View
Eden AI	deepinfra/deepseek-ai/DeepSeek-R1-0528	163,840	32,768	— / —	—	—	—	View
Hugging Face	deepseek-ai/DeepSeek-R1-0528	163,840	32,768	— / —	—	—	—	View
IO.NET	deepseek-ai/DeepSeek-R1-0528	128,000	4,096	— / —	—	—	—	View
Jiekou.AI	deepseek/deepseek-r1-0528	163,840	32,768	— / —	—	—	—	View
Kilo Gateway	deepseek/deepseek-r1-0528	163,840	32,768	— / —	—	—	—	View
Meganova	deepseek-ai/DeepSeek-R1-0528	163,840	64,000	— / —	—	—	—	View
Merge Gateway	deepseek/deepseek-r1-0528	65,536	32,768	— / —	—	—	—	View
NanoGPT	deepseek-ai/DeepSeek-R1-0528	163,840	32,768	— / —	—	—	—	View
Novita AI	deepseek/deepseek-r1-0528	163,840	32,768	— / —	—	—	—	View
DeepInfravia OpenRouter	deepseek/deepseek-r1-0528 fp4	163,840	32,768	$0.50 / $2.15	Yes	No	Yes	View
Novitavia OpenRouter	deepseek/deepseek-r1-0528 fp8	163,840	32,768	$0.70 / $2.50	Yes	Yes	No	View
SiliconFlowvia OpenRouter	deepseek/deepseek-r1-0528 fp8	163,840	147,456	$0.50 / $2.18	Yes	Yes	Yes	View
StreamLakevia OpenRouter	deepseek/deepseek-r1-0528 unknown	128,000	32,000	$0.571 / $2.286	Yes	No	Yes	View
Qiniu	deepseek-r1-0528	128,000	32,000	— / —	—	—	—	View
SiliconFlow	deepseek-ai/DeepSeek-R1	164,000	164,000	— / —	—	—	—	View
SiliconFlow (China)	deepseek-ai/DeepSeek-R1	164,000	164,000	— / —	—	—	—	View
SiliconFlow (China)	Pro/deepseek-ai/DeepSeek-R1	164,000	164,000	— / —	—	—	—	View
submodel	deepseek-ai/DeepSeek-R1-0528	75,000	163,840	— / —	—	—	—	View
TensorX	deepseek/deepseek-r1-0528	164,000	8,192	— / —	—	—	—	View

Prices and limits apply to each offer. A dash means the value is not provided. OpenRouter offers show the hosting provider and routing channel separately.

Pricing

Compare token rates and cache charges across available routes and hosting configurations.

USD per million tokens

Input from$0.50Per 1M input tokens

Output from$2.15Per 1M output tokens

Available offers21Across 16 providers

Lowest listed input and output prices across the offers above; they may belong to different providers. These are base rates, before conditional pricing, discounts or additional fees.

Endpoint price comparisonLatest observed snapshot · 4 lowest-priced configurations

InputOutput

DeepInfradeepinfra/fp4

$0.50$2.15

SiliconFlowsiliconflow/fp8

$0.50$2.18

StreamLakestreamlake

$0.571$2.286

Novitanovita/fp8

$0.70$2.50

deepseek/deepseek-r1-0528Observed Oct 10, 2026, 11:31 AM UTC

Input: $0.50
Output: $2.15
Cache read: $0.35

Hosting prices through OpenRouter All endpoints and cache rates

Endpoint	Input / 1M	Output / 1M	Cache read / 1M	Cache write / 1M
DeepInfradeepinfra/fp4	$0.50	$2.15	$0.35	—
Novitanovita/fp8	$0.70	$2.50	$0.35	—
SiliconFlowsiliconflow/fp8	$0.50	$2.18	—	—
StreamLakestreamlake	$0.571	$2.286	—	—

Each row is a hosting configuration. Open an offer in Providers for its full pricing conditions and additional fees. Endpoint observations may differ from the routing catalog's base price.

Performance

Review reported latency and output-throughput percentiles from the latest observed rolling window.

Latest 30-minute observation

Time to first token · milliseconds

Latency distributionRolling 30-minute snapshot · 4 endpoints

P50P75P90P99

DeepInfradeepinfra/fp4

664–2,860.12 ms

Novitanovita/fp8

1,184.5–6,305.18 ms

SiliconFlowsiliconflow/fp8

1,818–11,655.5 ms

StreamLakestreamlake

3,011–10,810.96 ms

All latency percentiles 4 endpoints

Endpoint	P50	P75	P90	P99
DeepInfradeepinfra/fp4	664 ms	912.25 ms	1,279.5 ms	2,860.12 ms
Novitanovita/fp8	1,184.5 ms	1,641.5 ms	2,482.8 ms	6,305.18 ms
SiliconFlowsiliconflow/fp8	1,818 ms	2,553.5 ms	5,065 ms	11,655.5 ms
StreamLakestreamlake	3,011 ms	3,945 ms	5,139.8 ms	10,810.96 ms

Output throughput · tokens per second

Throughput distributionRolling 30-minute snapshot · 4 endpoints

P50P75P90P99

DeepInfradeepinfra/fp4

31–48 t/s

Novitanovita/fp8

27–43.13 t/s

SiliconFlowsiliconflow/fp8

20–24 t/s

StreamLakestreamlake

48–70.63 t/s

All throughput percentiles 4 endpoints

Endpoint	P50	P75	P90	P99
DeepInfradeepinfra/fp4	31 t/s	36 t/s	40 t/s	48 t/s
Novitanovita/fp8	27 t/s	32 t/s	37 t/s	43.13 t/s
SiliconFlowsiliconflow/fp8	20 t/s	22 t/s	23 t/s	24 t/s
StreamLakestreamlake	48 t/s	53 t/s	63 t/s	70.63 t/s

Percentiles describe the observed request distribution. These are snapshots of a rolling window, not a historical time series.

Uptime

Compare successful request rates across the latest 5-minute, 30-minute, and 24-hour observations.

Successful requests by window

Availability by endpointCurrent overlapping windows · 4 endpoints

Endpoint5 minutes30 minutes24 hours

DeepInfradeepinfra/fp4Observed Oct 10, 2026, 11:31 AM UTC

100%

99.86%

Novitanovita/fp8Observed Oct 10, 2026, 11:31 AM UTC

100%

99.92%

SiliconFlowsiliconflow/fp8Observed Oct 10, 2026, 11:31 AM UTC

100%

96%

99.46%

StreamLakestreamlakeObserved Oct 10, 2026, 11:31 AM UTC

100%

99.32%

Reported by OpenRouter at the time of collection; rate-limited requests are excluded. A dash means no measurement was reported. This is not a live availability check.

Benchmarks

Review attributed evaluation results on each source's original scale.

External evaluations

deepseek/deepseek-r1-0528Observed Oct 10, 2026, 11:31 AM UTC

Benchmark profileArtificial Analysis · shared source scale from 0 to 20

Intelligence

13.1

Design Arena	Category	Elo	Win rate	Rank
models	3d	1,116	53.3%	79
models	codecategories	1,142	52.6%	83
models	dataviz	1,185	60.9%	58
models	gamedev	1,105	49.3%	87
models	svg	1,038	48.7%	74
models	uicomponent	1,106	54.9%	86
models	website	1,154	52.7%	83

Sources: Artificial Analysis and Design Arena, via OpenRouter. Scores retain their original scales and are not OpenAI Suite ratings.

Detailed evaluations

Evaluation scoresAccuracy or item F1 · higher is better

tau bench verified airlineDeepSeek: R1 0528

49.952%

gpqa diamondDeepSeek: R1 0528

77.189%

Evaluation	Configuration	Source	Results	Details
models · gamedev	DeepSeek-R1-0528deepseek/deepseek-r1-0528	Design Arena	1,097 Elo	Details
models · website	DeepSeek-R1-0528deepseek/deepseek-r1-0528	Design Arena	1,149 Elo	Details
models · codecategories	DeepSeek-R1-0528deepseek/deepseek-r1-0528	Design Arena	1,139 Elo	Details
tau bench verified airline	DeepSeek: R1 0528deepseek/deepseek-r1-0528	OpenRouter	49.952%	Details
models · 3d	DeepSeek-R1-0528deepseek/deepseek-r1-0528	Design Arena	1,111 Elo	Details
models · uicomponent	DeepSeek-R1-0528deepseek/deepseek-r1-0528	Design Arena	1,100 Elo	Details
models · svg	DeepSeek-R1-0528deepseek/deepseek-r1-0528	Design Arena	1,039 Elo	Details
gpqa diamond	DeepSeek: R1 0528deepseek/deepseek-r1-0528	OpenRouter	77.189%	Details
models · dataviz	DeepSeek-R1-0528deepseek/deepseek-r1-0528	Design Arena	1,182 Elo	Details
Composite indices	DeepSeek R1 0528 (May '25)deepseek/deepseek-r1-0528	Artificial Analysis	Intelligence 13.1 · Coding — · Agentic —	Details

Sources: Artificial Analysis, Design Arena and OpenRouter. Evaluation configurations and scales differ; results are not interchangeable. Updated Oct 23, 2026.

Technical details & API

Explore specifications, supported parameters, reasoning controls, and API identifiers for each OpenRouter route.

Current route specifications

deepseek/deepseek-r1-0528Observed Oct 10, 2026, 11:31 AM UTC

Context: 163,840
Maximum output: 32,768
Tokenizer: DeepSeek
Input: text
Output: text
Moderation: No
Instruction format: deepseek-r1
Knowledge cutoff: 2025-03-31
Added to OpenRouter: 2025-05-28

Base URLhttps://openrouter.ai/api/v1

Model IDdeepseek/deepseek-r1-0528

Version IDdeepseek/deepseek-r1-0528

Weights IDdeepseek-ai/DeepSeek-R1-0528

Reasoning

Required: Yes
Enabled by default: Not reported
Default effort: —
Supported efforts: —

Supported parameters

frequency_penalty
include_reasoning
logit_bias
logprobs
max_tokens
min_p
presence_penalty
reasoning
repetition_penalty
response_format
seed
stop
structured_outputs
temperature
tool_choice
tools
top_k
top_logprobs
top_p

Frequently asked questions

Answers based on the model metadata and serving offers currently in the catalog.

How much does DeepSeek R1 0528 cost?

Listed rates start at $0.50 per million input tokens and $2.15 per million output tokens. Prices vary by provider and configuration; see Pricing for details.

Which providers offer DeepSeek R1 0528?

There are 21 cataloged offers across 16 providers. Compare model IDs, prices and limits in Providers.

What is the context window?

The cataloged model context is 163,840 tokens. Each serving endpoint may apply a different limit.

Does it support tools and structured output?

Tool calling: Not reported. Structured output: Not reported. Support can vary by endpoint.

Resources

MIT

Hugging Face ↗

Catalog history · 2 recorded revisions

Metadata observations since this model was first cataloged. These are not model release versions.

Oct 20, 2026 · 97e9c4be5533
Oct 10, 2026 · 7f67a999c4cc