Text embedding API prices, per model

Model	Provider	Input /1M	Dimensions	Context	Released
text-embedding-3-small OpenAI Embedding OpenAI's smaller embedding model. Native 1536 dimensions, truncatable via the dimensions parameter down to 512. $0.02/—I/O 8K ctx Jan 2024 View model	OpenAI	$0.02	1536	8K	Jan 2024
V voyage-4-lite Voyage AI Embedding Voyage AI's entry tier in the voyage-4 family. Matryoshka dimensions, default 1024. $0.02/—I/O 32K ctx View model	Voyage AI	$0.02	1024	32K	—
V voyage-3.5-lite Voyage AI legacy Embedding Previous-generation cheap Voyage embedding model, superseded by voyage-4-lite. $0.02/—I/O 32K ctx View model	Voyage AI	$0.02	1024	32K	—
V voyage-4 Voyage AI Embedding Voyage AI's balanced general embedding model. Matryoshka dimensions, default 1024. $0.06/—I/O 32K ctx View model	Voyage AI	$0.06	1024	32K	—
V voyage-3.5 Voyage AI legacy Embedding Previous-generation Voyage embedding model, superseded by voyage-4. $0.06/—I/O 32K ctx View model	Voyage AI	$0.06	1024	32K	—
text-embedding-ada-002 OpenAI legacy Embedding OpenAI's original second-generation embedding model, superseded by the text-embedding-3 family. $0.1/—I/O 8K ctx Dec 2022 View model	OpenAI	$0.1	1536	8K	Dec 2022
Embed English v3 Cohere Embedding Cohere's English embedding model (embed-english-v3.0), 1024 dimensions. $0.1/—I/O 512 ctx View model	Cohere	$0.1	1024	512	—
Embed Multilingual v3 Cohere Embedding Cohere's multilingual embedding model (embed-multilingual-v3.0), 1024 dimensions across 100+ languages. $0.1/—I/O 512 ctx View model	Cohere	$0.1	1024	512	—
Mistral Embed Mistral Embedding Mistral's general embedding model (mistral-embed-2312), 1024 dimensions. $0.1/—I/O 8K ctx View model	Mistral	$0.1	1024	8K	—
Embed 4 Cohere Embedding Cohere's multimodal embedding model (embed-v4.0). Matryoshka dimensions of 256/512/1024/1536; the figure is the text-token rate. $0.12/—I/O 128K ctx View model	Cohere	$0.12	1536	128K	—
V voyage-4-large Voyage AI Embedding Voyage AI's largest general embedding model in the voyage-4 family. Matryoshka dimensions of 256/512/1024/2048, default 1024. $0.12/—I/O 32K ctx View model	Voyage AI	$0.12	1024	32K	—
text-embedding-3-large OpenAI Embedding OpenAI's larger embedding model. Native 3072 dimensions, truncatable down to 256. $0.13/—I/O 8K ctx Jan 2024 View model	OpenAI	$0.13	3072	8K	Jan 2024
gemini-embedding-001 Google Embedding Google's Gemini embedding model. Native 3072 dimensions with Matryoshka truncation (e.g. 1536/768/256). $0.15/—I/O 2K ctx View model	Google	$0.15	3072	2K	—
Codestral Embed Mistral Embedding Mistral's code-specialised embedding model (codestral-embed-2505). Native 3072 dimensions, default 1536, MRL-truncatable. $0.15/—I/O 8K ctx View model	Mistral	$0.15	3072	8K	—
V voyage-3-large Voyage AI legacy Embedding Previous-generation Voyage general embedding model, still listed alongside the voyage-4 family. $0.18/—I/O 32K ctx View model	Voyage AI	$0.18	1024	32K	—
V voyage-code-3 Voyage AI Embedding Voyage AI's code-specialised embedding model. Matryoshka dimensions, default 1024. $0.18/—I/O 32K ctx View model	Voyage AI	$0.18	1024	32K	—
V voyage-context-3 Voyage AI Embedding Voyage AI's contextualised-chunk embedding model. Matryoshka dimensions, default 1024. $0.18/—I/O 32K ctx View model	Voyage AI	$0.18	1024	32K	—
gemini-embedding-2 Google Embedding Google's next Gemini embedding model. Its price is primary-confirmed; dimensions are not yet confirmed on a primary source. $0.2/—I/O 2K ctx View model	Google	$0.2	—	2K	—

Embedding API pricing, per 1M input tokens.