Skip to content
Model infoModelsSnapshot: 15 Oct 2026

ByteDance / Seed

Doubao Seed 2.1 Lite (260915)

Volcengine-native Seed 2.1 Lite (260915) for reasoning, coding and agent workflows.

Language models (LLMs)
availableClosed weightsText generation
Official documentation (opens in a new tab)Reviewed 15 Oct 2026
12 of 20 specification fields collected
Input token limit1024KSee specification notes below
Maximum output256KSee specification notes below
Input / 1M tokensCN¥0.80Volcengine Ark (China)
Output / 1M tokensCN¥2.70CNY · billing conditions below

Specifications

Provider-documented facts. “Not collected” means we have not verified that field yet, not that the feature is unsupported.

Context / input limit
1024KSource for Context / input limit (opens in a new tab)

Provider K notation retained; shared context is not an independent budget for each input/output component.

Maximum input tokens
1024KSource for Maximum input tokens (opens in a new tab)
Maximum output
256KSource for Maximum output (opens in a new tab)

Published answer limit. Shared context and reasoning consumption also constrain a request.

Reasoning
SupportedSource for Reasoning (opens in a new tab)

The catalog states models without the deep-thinking tag do not support deep thinking.

Tool calling
SupportedSource for Tool calling (opens in a new tab)
Structured output
SupportedSource for Structured output (opens in a new tab)

API schema mode restrictions can change when explicit caching is enabled.

Streaming
Not collected
Batch processing
Not collected
Fine-tuning
Not collected
Knowledge cutoff
Not collected
Release date
Not collected
Parameters
Not collected
Architecture
Not collected
Languages
Not collected
Context caching
SupportedSource for Context caching (opens in a new tab)

Implicit and explicit cache modes are documented for this selector.

Published chain-of-thought limit
256KSource for Published chain-of-thought limit (opens in a new tab)

Published reasoning allowance, not additional space beyond context.

Published request quota
Max RPM: 30000 Max TPM: 5000000Source for Published request quota (opens in a new tab)

Published quotas are not rigid guarantees; account, endpoint and platform load apply.

Input modalities
Text, Image, Video, AudioSource for Input modalities (opens in a new tab)
Output modalities
TextSource for Output modalities (opens in a new tab)
Release month
2026-09Source for Release month (opens in a new tab)

Release-log month only; no exact day inferred from the API suffix.

Access & pricing

Prices belong to an endpoint and billing tier. A provider listing does not guarantee access for every account or region.

Volcengine Ark (China) · Standard online API; other service modes have separately published family tariffs

Provider source (opens in a new tab)
API model ID
doubao-seed-2-1-lite-260915
Access
Native hosted API requires Volcengine service activation, an eligible account and API key. No inference was executed.
Regions
China (Beijing)
Service tier
Standard online API; other service modes have separately published family tariffsSource for Service tier (opens in a new tab)
Published request quota
Max RPM: 30000 Max TPM: 5000000Source for Published request quota (opens in a new tab)

Published quotas are not rigid guarantees; account, endpoint and platform load apply.

Input modalities
Text, Image, Video, AudioSource for Input modalities (opens in a new tab)
Output modalities
TextSource for Output modalities (opens in a new tab)
Context / input limit
1024KSource for Context / input limit (opens in a new tab)

Provider K notation retained; shared context is not an independent budget for each input/output component.

Maximum input tokens
1024KSource for Maximum input tokens (opens in a new tab)
Maximum output
256KSource for Maximum output (opens in a new tab)

Published answer limit. Shared context and reasoning consumption also constrain a request.

Prices retain their published currencies and billing units below.

Published billing units · original currency · observed 15 Oct 2026
CategoryUsageRateUnit
Volcengine Standard online family tariffNon-audio input · Input length [0, 1024] (thousand tokens)1CN¥0.801M tokens
Volcengine Standard online family tariffAudio input · Input length [0, 1024] (thousand tokens)1CN¥12.001M tokens
Volcengine Standard online family tariffExplicit cache storage · Input length [0, 1024] (thousand tokens)1CN¥0.0171M tokens/hour
Volcengine Standard online family tariffCached non-audio input · Input length [0, 1024] (thousand tokens)1CN¥0.161M tokens
Volcengine Standard online family tariffCached audio input · Input length [0, 1024] (thousand tokens)1CN¥2.401M tokens
Volcengine Standard online family tariffOutput · Input length [0, 1024] (thousand tokens)1CN¥2.701M tokens
Volcengine Low-latency online family tariffNon-audio input · Input length [0, 1024] (thousand tokens)2CN¥1.601M tokens
Volcengine Low-latency online family tariffAudio input · Input length [0, 1024] (thousand tokens)2CN¥24.001M tokens
Volcengine Low-latency online family tariffCached non-audio input · Input length [0, 1024] (thousand tokens)2CN¥0.321M tokens
Volcengine Low-latency online family tariffCached audio input · Input length [0, 1024] (thousand tokens)2CN¥4.801M tokens
Volcengine Low-latency online family tariffOutput · Input length [0, 1024] (thousand tokens)2CN¥5.401M tokens
Volcengine Flex online family tariffNon-audio input · Input length [0, 1024] (thousand tokens)3CN¥0.401M tokens
Volcengine Flex online family tariffAudio input · Input length [0, 1024] (thousand tokens)3CN¥6.001M tokens
Volcengine Flex online family tariffCached non-audio input · Input length [0, 1024] (thousand tokens)3CN¥0.161M tokens
Volcengine Flex online family tariffCached audio input · Input length [0, 1024] (thousand tokens)3CN¥2.401M tokens
Volcengine Flex online family tariffOutput · Input length [0, 1024] (thousand tokens)3CN¥1.351M tokens
Volcengine Batch inference family tariffNon-audio input · Input length [0, 1024] (thousand tokens)4CN¥0.401M tokens
Volcengine Batch inference family tariffAudio input · Input length [0, 1024] (thousand tokens)4CN¥6.001M tokens
Volcengine Batch inference family tariffCached non-audio input · Input length [0, 1024] (thousand tokens)4CN¥0.161M tokens
Volcengine Batch inference family tariffCached audio input · Input length [0, 1024] (thousand tokens)4CN¥2.401M tokens
Volcengine Batch inference family tariffOutput · Input length [0, 1024] (thousand tokens)4CN¥1.351M tokens

Rate conditions

  1. Published family tariff for doubao-seed-2.1-lite, Standard online, prompt length [0, 1024] thousand tokens. Applicable request length selects the whole-request rate, not incremental blocks. Audio input is separately priced. This tariff does not itself prove every service mode is enabled on every dated endpoint. Explicit cache storage is separate; Flex permits implicit cache only. Native currency retained; no exchange-rate conversion.
  2. Published family tariff for doubao-seed-2.1-lite, Low-latency online, prompt length [0, 1024] thousand tokens. Applicable request length selects the whole-request rate, not incremental blocks. Audio input is separately priced. This tariff does not itself prove every service mode is enabled on every dated endpoint. Explicit cache storage is separate; Flex permits implicit cache only. Native currency retained; no exchange-rate conversion.
  3. Published family tariff for doubao-seed-2.1-lite, Flex online, prompt length [0, 1024] thousand tokens. Applicable request length selects the whole-request rate, not incremental blocks. Audio input is separately priced. This tariff does not itself prove every service mode is enabled on every dated endpoint. Explicit cache storage is separate; Flex permits implicit cache only. Native currency retained; no exchange-rate conversion.
  4. Published family tariff for doubao-seed-2.1-lite, Batch inference, prompt length [0, 1024] thousand tokens. Applicable request length selects the whole-request rate, not incremental blocks. Audio input is separately priced. This tariff does not itself prove every service mode is enabled on every dated endpoint. Explicit cache storage is separate; Flex permits implicit cache only. Native currency retained; no exchange-rate conversion.
Pricing source (opens in a new tab)

Data & enterprise terms

Retention, training use, regional processing and enterprise terms have not been summarized for this endpoint.

Caching

Caching can reuse earlier input processing. Availability, lifetime and billing depend on the endpoint. A cache hit still generates a new response.

01 / INITIAL INPUTProcess & retain

The system processes eligible input and retains reusable state.

02 / LATER REQUESTCheck available state

Reuse depends on matching input, session state and the cache policy.

03 / CACHE HITReuse & generate

Reusing state can reduce repeated work. Any charges follow the billing rules below.

Volcengine Ark (China) · Standard online API; other service modes have separately published family tariffs

doubao-seed-2-1-lite-260915
Documented
Support evidence (opens in a new tab)
How it works
Automatic implicit prefix reuse; explicit Responses API prefix/session cachingSource for How it works (opens in a new tab)
Minimum reusable input
Implicit: 1024 input tokens by default. Explicit prefix: at least 256 tokens; session cache: no minimum.Source for Minimum reusable input (opens in a new tab)
Lifetime & refresh
Implicit cache lifetime is dynamic. Explicit expire_at may be at most 7 days ahead and is not refreshed by reuse.Source for Lifetime & refresh (opens in a new tab)
How to enable it
Explicit: store=true, caching.type=enabled and previous_response_id. Keep all earlier cached turns enabled.Source for How to enable it (opens in a new tab)
Usage fields
Reuse an unchanged prefix or earlier Responses conversation. Implicit hits are not guaranteed.Source for Usage fields (opens in a new tab)
Billing rules
Implicit storage is free. Explicit storage is billed separately from cached-input and output tokens, rounded up to a whole natural hour. See the scoped CNY rate table for amounts.Source for Billing rules (opens in a new tab)
Limits & data boundaries
Explicit cache history permits json_object, not json_schema. Keep thinking settings consistent, leave instructions empty and set tools only on the first turn. Prefix creation cannot stream. Model-version switching can temporarily miss caches while storage is still billed.Source for Limits & data boundaries (opens in a new tab)
Cache reads, writes & storage
ChargeApplies toRate / unit
Explicit cache storage · Input length [0, 1024] (thousand tokens)1Volcengine Standard online family tariffCN¥0.017 / 1M tokens/hourPricing evidence (opens in a new tab)
Cached non-audio input · Input length [0, 1024] (thousand tokens)1Volcengine Standard online family tariffCN¥0.16 / 1M tokensPricing evidence (opens in a new tab)
Cached audio input · Input length [0, 1024] (thousand tokens)1Volcengine Standard online family tariffCN¥2.40 / 1M tokensPricing evidence (opens in a new tab)
Cached non-audio input · Input length [0, 1024] (thousand tokens)2Volcengine Low-latency online family tariffCN¥0.32 / 1M tokensPricing evidence (opens in a new tab)
Cached audio input · Input length [0, 1024] (thousand tokens)2Volcengine Low-latency online family tariffCN¥4.80 / 1M tokensPricing evidence (opens in a new tab)
Cached non-audio input · Input length [0, 1024] (thousand tokens)3Volcengine Flex online family tariffCN¥0.16 / 1M tokensPricing evidence (opens in a new tab)
Cached audio input · Input length [0, 1024] (thousand tokens)3Volcengine Flex online family tariffCN¥2.40 / 1M tokensPricing evidence (opens in a new tab)
Cached non-audio input · Input length [0, 1024] (thousand tokens)4Volcengine Batch inference family tariffCN¥0.16 / 1M tokensPricing evidence (opens in a new tab)
Cached audio input · Input length [0, 1024] (thousand tokens)4Volcengine Batch inference family tariffCN¥2.40 / 1M tokensPricing evidence (opens in a new tab)

Rate conditions

  1. Published family tariff for doubao-seed-2.1-lite, Standard online, prompt length [0, 1024] thousand tokens. Applicable request length selects the whole-request rate, not incremental blocks. Audio input is separately priced. This tariff does not itself prove every service mode is enabled on every dated endpoint. Explicit cache storage is separate; Flex permits implicit cache only. Native currency retained; no exchange-rate conversion.
  2. Published family tariff for doubao-seed-2.1-lite, Low-latency online, prompt length [0, 1024] thousand tokens. Applicable request length selects the whole-request rate, not incremental blocks. Audio input is separately priced. This tariff does not itself prove every service mode is enabled on every dated endpoint. Explicit cache storage is separate; Flex permits implicit cache only. Native currency retained; no exchange-rate conversion.
  3. Published family tariff for doubao-seed-2.1-lite, Flex online, prompt length [0, 1024] thousand tokens. Applicable request length selects the whole-request rate, not incremental blocks. Audio input is separately priced. This tariff does not itself prove every service mode is enabled on every dated endpoint. Explicit cache storage is separate; Flex permits implicit cache only. Native currency retained; no exchange-rate conversion.
  4. Published family tariff for doubao-seed-2.1-lite, Batch inference, prompt length [0, 1024] thousand tokens. Applicable request length selects the whole-request rate, not incremental blocks. Audio input is separately priced. This tariff does not itself prove every service mode is enabled on every dated endpoint. Explicit cache storage is separate; Flex permits implicit cache only. Native currency retained; no exchange-rate conversion.

Observed 15 Oct 2026. Write rates and storage fees are different charges; use the billing rules above.

Caching reviewed 15 Oct 2026

What the measurements say

No benchmark observations have been added for this model. We do not turn a vendor description into a performance score.

Latency and throughput depend on the host, request size, reasoning settings and load. There is no measured speed or global usage figure in this record.

The record behind the model

Why it is included

Inferred from the named native hosted foundation-model offering, official release history and current ByteDance/ByteDance-Seed public inventories. No matching public checkpoint was found. This is not an explicit published weight-license declaration.

Classification evidence (opens in a new tab)

Sources reviewed