Search papers, labs, and topics across Lattice.
The paper introduces Giga-Embeddings, a suite of text embedding models that leverage a Mixture-of-Experts architecture to optimize retrieval quality and serving efficiency. The largest model, featuring 10 billion parameters with 1.8 billion active parameters per token, achieves superior performance across multiple benchmarks, including English and multilingual tasks, while processing 114.5k tokens per second—25% faster than a dense 3B model. Additionally, a compact 480M model is trained with a dimension-agnostic objective, outperforming existing models like FRIDA with fewer parameters, showcasing the versatility and efficiency of the Giga-Embeddings family.
Achieving 25% higher throughput than dense models, Giga-Embeddings redefines efficiency in high-quality text embedding generation.
We introduce Giga-Embeddings, a family of text embedding models designed to combine strong retrieval quality with efficient serving. Its largest member is a sparse 10B-parameter Mixture-of-Experts encoder with approximately 1.8B active parameters per token. Across English, Russian, multilingual, and code MTEB benchmarks, this model achieves the strongest aggregate performance within the family on all four evaluated suites. In our vLLM benchmark with 1024-token inputs, it processes 114.5k tokens per second, providing 25 percent higher throughput than the dense 3B model and 1.56-2.65x the throughput of the evaluated external systems. The family also includes a dense 3B encoder and a distilled 480M encoder for tighter compute and memory budgets. We train the compact model using a dimension-agnostic objective that aligns teacher and student similarity distributions. The resulting 480M model scores 70.98 on Russian MTEB, surpassing FRIDA while using 42 percent fewer parameters. We release all three model checkpoints.
