Voyage AI embedding model
voyage-4
Voyage 4
Voyage 4 is the middle tier of Voyage AI's Voyage 4 embedding series, for general-purpose and multilingual retrieval. it sits between voyage-4-lite and voyage-4-large, and embeddings from all three can be used interchangeably.
01
price
per 1M input tokens
embeddings bill input only — there are no output tokens. octomind bills the listed provider price, with no markup, on every plan, including the free one (pay as you go from credits).
02
what real work costs
| workload | cost |
|---|---|
| index 10,000 help articles ~800 tokens each, 8M tokens | $0.48 |
| index a 1M-line codebase ~10 tokens a line, 10M tokens | $0.60 |
| embed 1M search queries ~20 tokens each, 20M tokens | $1.20 |
at the listed price. re-embedding after a model change costs the same again.
03
specs
- provider
- Voyage AI
- max input
- about 32k tokens per text
- dimensions
- 1,024 by default; also 256, 512 and 2,048
- storage per 1M vectors
- 4.1 GB at 1,024 dims · 2 GB at 512 dims · 1 GB at 256 dims (float32)
- accepts
- text
- multilingual
- yes
- weights
- closed
- released
- January 15, 2026
- model id
voyage-4
integration notes
- a request takes at most 1,000 texts and 320K total tokens.
- inputs over 32,000 tokens are silently truncated by default; set truncation to false to get an error instead.
- input_type query or document prepends a fixed retrieval prompt; omitting it embeds the raw text.
04
when to pick it
pick it if
- you want half voyage-4-large's price; Voyage reports large 1.87% ahead on average
- you need multilingual retrieval or 256-2048 dimension embeddings
- you may mix it with other Voyage 4 models in one index
skip it if
- voyage-4-lite is accurate enough at a third of the price
- you are indexing source code; voyage-code-4 is optimized for code retrieval
05
where it sits on price
7th cheapest of the 10 embedding models in the catalog.
06
use Voyage 4 in octomind
octocode, octomind's semantic code search, takes any octohub embedding model by id. one octohub key covers it.
[embedding]
code_model = "octohub:voyage-4"
text_model = "octohub:voyage-4"07
compare Voyage 4
08
similar embeddings
nearest embedding models by price.
- Jina v5 Omni Small $0.05/M input
- Jina v5 Text Small $0.05/M input
- Jina v5 Omni Nano $0.02/M input
- Jina v5 Text Nano $0.02/M input
09
Voyage 4 faq
how much does voyage-4 cost?
$0.06/M input tokens; embeddings have no output charge. index a 1M-line codebase (~10 tokens a line, 10M tokens) costs about $0.60.
how many dimensions does voyage-4 have?
1,024 by default; also 256, 512 and 2,048. at the default size, one million float32 vectors take 4.1 GB.
what is the maximum input length of Voyage 4?
about 32k tokens per text (≈ 50 pages of prose or 3,000 lines of code). longer documents need chunking before embedding.
how do i use Voyage 4 with octocode?
set code_model and text_model to "octohub:voyage-4" in the [embedding] section of octocode's config.toml, then re-index.
use Voyage 4 in octomind
use it for semantic code search with octocode, or call it through octohub.