Voyage Multimodal 3
All-in-one embedding model for interleaved text, images, and screenshots in a shared vector space.
About
Voyage Multimodal 3 is Voyage AI's embedding model for interleaved text and images, released November 12, 2024. A query mixing screenshots, product photos, and text is fused into a single 1024-dimensional vector in a shared space, which makes it a natural fit for multimodal RAG: retrieve the right chart, UI screenshot, or document page with one embedding call instead of separate text and image pipelines.
Voyage publishes an evaluation spreadsheet claiming leads over alternatives including Cohere's multimodal embeddings, but those are provider-run numbers. The model is API-only with no published weights, and Voyage offers the first 200 million tokens free. A preview successor, voyage-multimodal-3.5, adds video and flexible output dimensions, but 3 remains the stable general-availability multimodal embedding.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
No independent public benchmark table found; Voyage claims leading results over Cohere multimodal v3 and others in its published evaluation spreadsheet (provider-reported)
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: real Voyage AI release Nov 12, 2024 with a public launch post and eval spreadsheet; integrated across LangChain4j, Haystack, and Symfony AI docs showing genuine developer adoption. Closed weights. Strengths: purpose-built interleaved text+image embeddings, generous 200M-token free tier. Docked for: (1) retrieval-quality claims are provider-run with no independent benchmark found; (2) closed weights and API-only access.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 7.0/10
- Maintenance · recency, releases, issue load
- 7.0/10
- Openness · license + docs
- 2.0/10
- Manual review · sanity read + red-flag check
- 7.0/10
- · Voyage AI release 2024-11-12; 1024-dim interleaved text+image embeddings
- · Integrated in LangChain4j, Haystack, Symfony AI docs — real developer adoption
- · First 200M tokens free; 3.5 preview adds video
- · Retrieval-quality claims are provider-run; closed weights, API-only