Match each 2026 embedding vendor to a defining feature or pricing fact

Each major embedding vendor in 2026 has a defining selling point, and recognizing them is faster than reading every model card. OpenAI text-embedding-3-small is the cost-optimized default: $0.02 per million tokens, 1536 dimensions, Matryoshka-truncatable. Its sibling text-embedding-3-large is the premium quality option at $0.13 per million. Voyage voyage-code-3 is the code specialist, distinct from Voyage's general voyage-3 model. Code-aware tokenizer and code-heavy training mix give it a 10 to 30 percent edge on benchmarks like CodeSearchNet. BGE-M3 is the open-weights multilingual leader from BAAI. It handles 100+ languages and outputs three kinds of representations (dense, sparse, multi-vector) from a single forward pass. Cohere embed-v3 and v4 are tightly paired with the rerank-3 cross-encoder, which makes them the popular choice for production RAG stacks that need both a first-stage embedder and a reranker. Nomic Embed v1.5 is the open-weights self-host pick. Matryoshka-trained, deployable on a single GPU, no API egress.

Each vendor in the 2026 lineup has a distinct positioning that drives selection. **OpenAI text-embedding-3-small** is the cost-optimized default at $0.02 per million tokens. 1536 dimensions, Matryoshka-truncatable, hosted, low operational overhead. The 'just pick this' answer for general English RAG when you do not want to think about model selection. **Voyage voyage-code-3** is the code specialist. Code-aware tokenization plus a code-heavy training mix give it 10 to 30 percent better recall than general embedders on CodeSearchNet-style benchmarks. Used by Cursor, Continue, and similar code-aware products. **BGE-M3** is the open-weights multilingual leader from BAAI. 100+ languages with a shared embedding space, plus hybrid output (dense, sparse, and multi-vector from one call). Self-hostable on a single H100. **Cohere embed-v3 and v4** ship tightly paired with rerank-3, a cross-encoder tuned to the embedder's domain. The pairing is the value: picking embedder and reranker from one vendor simplifies the stack and improves end to end quality. Nomic Embed v1.5 is the on-prem open-weights pick. Matryoshka-trained, strong MTEB, designed for the 'no API egress' constraint that enterprise and regulated deployments often have.