Skip to content
Model infoModelsSnapshot: 15 Oct 2026

Alibaba Cloud / Tongyi Text Embedding

Tongyi Text Embedding v3

Multilingual text retrieval with three vector sizes and query/document task modes.

availableClosed weightsEmbeddings
Official documentation (opens in a new tab)Reviewed 14 Oct 2026
24 of 31 specification fields collected
Input token limit8.2K8,192 tokens
Vector dimensions1024, 768, 512See specification notes below
Input billingSee ratesAlibaba Model Studio · Singapore
Output billingSee ratesPublished units and 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
Not collected
Maximum input tokens
8,192 tokensSource for Maximum input tokens (opens in a new tab)

Per input item, not an aggregate request token ceiling.

Maximum output
Not collected
Output modalities
EmbeddingsSource for Output modalities (opens in a new tab)

Native API result type takes precedence over generic model-card modality labels.

Reasoning
Not collected
Streaming
Not collected
Knowledge cutoff
Not collected
Release date
2024-07-12Source for Release date (opens in a new tab)

Earliest reviewed regional release; not inferred from a snapshot suffix.

Parameters
Not collected
Architecture
Not collected
Languages
Chinese, English, Spanish, French, Portuguese, Indonesian, Japanese, Korean, German, RussianSource for Languages (opens in a new tab)

Named examples from the provider, not an exhaustive 50+ language roster.

API aliases
text-embedding-v3Source for API aliases (opens in a new tab)

Only this exact selector is assigned. Dated variants are not assumed equivalent.

Data retention
Standard inference prompts/responses may be retained for up to 30 days with legal/security exceptions; stateful records/files remain until deleted. Exact-selector zero-data-retention eligibility has not been established in this review.Source for Data retention (opens in a new tab)

Original International policy reference. The domestic Beijing contract was not verified and is not inferred from this policy.

Embedding dimensions
1024, 768, 512Source for Embedding dimensions (opens in a new tab)

Model-specific dimensions table; default 1024. Generic parameter prose has inconsistent model qualifiers.

Items per embedding request
10Source for Items per embedding request (opens in a new tab)

Maximum items in one synchronous request; separate from asynchronous Batch Inference support.

Embedding output modes
dense, sparse, dense&sparseSource for Embedding output modes (opens in a new tab)

DashScope output_type; default dense. OpenAI-compatible API returns float dense vectors only.

Retrieval task modes
query, documentSource for Retrieval task modes (opens in a new tab)

DashScope text_type; default document. Use matching query/document representations for retrieval.

Request mode
OpenAI-compatible embeddings or DashScope text embedding; strings, string arrays or file inputSource for Request mode (opens in a new tab)
Language coverage
Provider claims 50+ major languagesSource for Language coverage (opens in a new tab)
Free usage allowance
500,000 tokens Valid for 90 days from the date of Model Studio activation, model release, or application approval, whichever is laterSource for Free usage allowance (opens in a new tab)

Original International Singapore quota. Regional account eligibility can differ.

Usage metering
Input-token billing; vector dimensions are not generated output tokens.Source for Usage metering (opens in a new tab)
Request quotas
6,000 RPM; 24,000,000 TPMSource for Request quotas (opens in a new tab)

Access & pricing

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

Alibaba Model Studio · Singapore

Provider source (opens in a new tab)
API model ID
text-embedding-v3
Access
https://{WorkspaceId}.ap-southeast-1.maas.aliyuncs.com/compatible-mode/v1/embeddings or https://{WorkspaceId}.ap-southeast-1.maas.aliyuncs.com/api/v1/services/embeddings/text-embedding/text-embedding. Use the matching regional workspace API key.
Regions
Singapore
Request quotas
6,000 RPM; 24,000,000 TPMSource for Request quotas (opens in a new tab)

Prices retain their published currencies and billing units below.

Published billing units · original currency · observed 14 Oct 2026
CategoryUsageRateUnit
EmbeddingsText input1$0.071M tokens
EmbeddingsText input2CN¥0.5141M tokens

Rate conditions

  1. Standard Singapore rate, excluding promotions. Input embeddings are not generated answers; no text-question estimate is inferred.
  2. Published domestic-documentation CNY tariff, not a currency conversion from USD. Singapore rate depends on the account billing currency. No generated-answer estimate is inferred.
Pricing source (opens in a new tab)Pricing source (opens in a new tab)

Data & enterprise terms

Standard inference prompts/responses may be retained for up to 30 days with legal/security exceptions; stateful records/files remain until deleted. Exact-selector zero-data-retention eligibility has not been established in this review.

Read the service policy (opens in a new tab)

Alibaba Model Studio · China (Beijing)

Provider source (opens in a new tab)
API model ID
text-embedding-v3
Access
https://{WorkspaceId}.cn-beijing.maas.aliyuncs.com/compatible-mode/v1/embeddings. Use a Beijing workspace API key. Text embedding also supports the DashScope text-embedding API.
Regions
China (Beijing)
Batch processing
Not collected

Card marks Batch Inference unsupported, but the same card and API guide publish a Batch File price. Confirm endpoint eligibility.

Request quotas
1,800 RPM; 1,200,000 TPMSource for Request quotas (opens in a new tab)
Embedding dimensions
1024, 768, 512, 256, 128, 64Source for Embedding dimensions (opens in a new tab)

Beijing documentation has additional dimensions not listed in the International Singapore table.

Prices retain their published currencies and billing units below.

Published billing units · original currency · observed 14 Oct 2026
CategoryUsageRateUnit
EmbeddingsText input1CN¥0.501M tokens
Batch embeddingsBatch text input2CN¥0.251M tokens

Rate conditions

  1. Published domestic-documentation CNY tariff, not a currency conversion from USD. Beijing workspace. No generated-answer estimate is inferred.
  2. Published domestic-documentation CNY tariff, not a currency conversion from USD. Beijing workspace. Batch availability is contradictory: the card marks it unsupported but publishes this rate. Confirm before use. No generated-answer estimate is inferred.
Pricing source (opens in a new tab)

Data & enterprise terms

Account-specific regional retention and enterprise policy require confirmation.

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.

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

Classification inference: the official GTE developer card explicitly distinguishes text-embedding-v1/v2/v3 commercial API models from the public models. Exact private checkpoint lineage and parameter counts are not disclosed, and public GTE weights are not treated as a verified release of this service.

Classification evidence (opens in a new tab)

Lifecycle dates

Platform release · Singapore12 Jul 2024Source (opens in a new tab)
Platform release · China (Beijing)12 Jul 2024Source (opens in a new tab)

Sources reviewed